Show
Ignore:
Timestamp:
08/30/09 22:13:15 (15 years ago)
Author:
smidl
Message:

doc

Files:
1 modified

Legend:

Unmodified
Added
Removed

  • library/doc/html/exp__family_8h_source.html

    r472 r591  
    1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> 
    2 <html><head><meta http-equiv="Content-Type" content="text/html;charset=UTF-8"> 
     1<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> 
     2<html xmlns="http://www.w3.org/1999/xhtml"> 
     3<head> 
     4<meta http-equiv="Content-Type" content="text/xhtml;charset=UTF-8"/> 
    35<title>mixpp: exp_family.h Source File</title> 
    4 <link href="tabs.css" rel="stylesheet" type="text/css"> 
    5 <link href="doxygen.css" rel="stylesheet" type="text/css"> 
    6 </head><body> 
    7 <!-- Generated by Doxygen 1.5.9 --> 
     6<link href="tabs.css" rel="stylesheet" type="text/css"/> 
     7<link href="doxygen.css" rel="stylesheet" type="text/css"/> 
     8</head> 
     9<body> 
     10<!-- Generated by Doxygen 1.6.1 --> 
    811<script type="text/javascript"> 
    912<!-- 
    … …  
    6770<a name="l00015"></a>00015 <span class="preprocessor"></span> 
    6871<a name="l00016"></a>00016  
    69 <a name="l00017"></a>00017 <span class="preprocessor">#include "../shared_ptr.h"</span> 
    70 <a name="l00018"></a>00018 <span class="preprocessor">#include "../base/bdmbase.h"</span> 
    71 <a name="l00019"></a>00019 <span class="preprocessor">#include "../math/chmat.h"</span> 
     72<a name="l00017"></a>00017 <span class="preprocessor">#include &quot;../shared_ptr.h&quot;</span> 
     73<a name="l00018"></a>00018 <span class="preprocessor">#include &quot;../base/bdmbase.h&quot;</span> 
     74<a name="l00019"></a>00019 <span class="preprocessor">#include &quot;../math/chmat.h&quot;</span> 
    7275<a name="l00020"></a>00020  
    7376<a name="l00021"></a>00021 <span class="keyword">namespace </span>bdm 
    … …  
    7578<a name="l00023"></a>00023  
    7679<a name="l00024"></a>00024  
    77 <a name="l00026"></a>00026         <span class="keyword">extern</span> Uniform_RNG UniRNG; 
    78 <a name="l00028"></a>00028         <span class="keyword">extern</span> Normal_RNG NorRNG; 
    79 <a name="l00030"></a>00030         <span class="keyword">extern</span> <a class="code" href="classitpp_1_1Gamma__RNG.html" title="Gamma distribution.">Gamma_RNG</a> GamRNG; 
     80<a name="l00026"></a>00026 <span class="keyword">extern</span> Uniform_RNG UniRNG; 
     81<a name="l00028"></a>00028 <span class="keyword">extern</span> Normal_RNG NorRNG; 
     82<a name="l00030"></a>00030 <span class="keyword">extern</span> Gamma_RNG GamRNG; 
    8083<a name="l00031"></a>00031  
    81 <a name="l00038"></a><a class="code" href="classbdm_1_1eEF.html">00038</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> 
    82 <a name="l00039"></a>00039         { 
    83 <a name="l00040"></a>00040                 <span class="keyword">public</span>: 
     84<a name="l00038"></a><a class="code" href="classbdm_1_1eEF.html">00038</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> 
     85<a name="l00039"></a>00039 { 
     86<a name="l00040"></a>00040         <span class="keyword">public</span>: 
    8487<a name="l00041"></a>00041 <span class="comment">//      eEF() :epdf() {};</span> 
    85 <a name="l00043"></a><a class="code" href="classbdm_1_1eEF.html#d5459d472d0feca7cf1fb5f65c4b9ef4">00043</a> <span class="comment"></span>                        <a class="code" href="classbdm_1_1eEF.html#d5459d472d0feca7cf1fb5f65c4b9ef4" title="default constructor">eEF</a> ( ) :<a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> ( ) {}; 
    86 <a name="l00045"></a>00045                         <span class="keyword">virtual</span> <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eEF.html#cd678fc9b02007a4b8d6692e746f1bf8" title="logarithm of the normalizing constant, ">lognc</a>() <span class="keyword">const</span> =0; 
    87 <a name="l00047"></a><a class="code" href="classbdm_1_1eEF.html#deef7d6273ba4d5a5cf0bbd91ec7277a">00047</a>                         <span class="keyword">virtual</span> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eEF.html#deef7d6273ba4d5a5cf0bbd91ec7277a" title="TODO decide if it is really needed.">dupdate</a> ( mat &amp;v ) {it_error ( <span class="stringliteral">"Not implemented"</span> );}; 
    88 <a name="l00049"></a><a class="code" href="classbdm_1_1eEF.html#a4135778ecd9ab774762936c82a097c6">00049</a>                         <span class="keyword">virtual</span> <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eEF.html#a4135778ecd9ab774762936c82a097c6" title="Evaluate normalized log-probability.">evallog_nn</a> ( <span class="keyword">const</span> vec &amp;val )<span class="keyword"> const</span>{it_error ( <span class="stringliteral">"Not implemented"</span> );<span class="keywordflow">return</span> 0.0;}; 
    89 <a name="l00051"></a><a class="code" href="classbdm_1_1eEF.html#a36d06ecdd6f4c79dc122510eaccc692">00051</a>                         <span class="keyword">virtual</span> <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eEF.html#a36d06ecdd6f4c79dc122510eaccc692" title="Evaluate normalized log-probability.">evallog</a> ( <span class="keyword">const</span> vec &amp;val )<span class="keyword"> const </span>{ 
    90 <a name="l00052"></a>00052                                 <span class="keywordtype">double</span> tmp; 
    91 <a name="l00053"></a>00053                                 tmp= <a class="code" href="classbdm_1_1eEF.html#a4135778ecd9ab774762936c82a097c6" title="Evaluate normalized log-probability.">evallog_nn</a> ( val )-<a class="code" href="classbdm_1_1eEF.html#cd678fc9b02007a4b8d6692e746f1bf8" title="logarithm of the normalizing constant, ">lognc</a>(); 
    92 <a name="l00054"></a>00054 <span class="comment">//                              it_assert_debug ( std::isfinite ( tmp ),"Infinite value" ); </span> 
    93 <a name="l00055"></a>00055                                 <span class="keywordflow">return</span> tmp;} 
    94 <a name="l00057"></a><a class="code" href="classbdm_1_1eEF.html#79a7c8ea8c02e45d410bd1d7ffd72b41">00057</a>                         <span class="keyword">virtual</span> vec <a class="code" href="classbdm_1_1eEF.html#a36d06ecdd6f4c79dc122510eaccc692" title="Evaluate normalized log-probability.">evallog</a> ( <span class="keyword">const</span> mat &amp;Val )<span class="keyword"> const</span> 
    95 <a name="l00058"></a>00058 <span class="keyword">                        </span>{ 
    96 <a name="l00059"></a>00059                                 vec x ( Val.cols() ); 
    97 <a name="l00060"></a>00060                                 <span class="keywordflow">for</span> ( <span class="keywordtype">int</span> i=0;i&lt;Val.cols();i++ ) {x ( i ) =<a class="code" href="classbdm_1_1eEF.html#a4135778ecd9ab774762936c82a097c6" title="Evaluate normalized log-probability.">evallog_nn</a> ( Val.get_col ( i ) ) ;} 
    98 <a name="l00061"></a>00061                                 <span class="keywordflow">return</span> x-<a class="code" href="classbdm_1_1eEF.html#cd678fc9b02007a4b8d6692e746f1bf8" title="logarithm of the normalizing constant, ">lognc</a>(); 
    99 <a name="l00062"></a>00062                         } 
    100 <a name="l00064"></a><a class="code" href="classbdm_1_1eEF.html#cf38af29e8e3d650c640509a52396053">00064</a>                         <span class="keyword">virtual</span> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eEF.html#cf38af29e8e3d650c640509a52396053" title="Power of the density, used e.g. to flatten the density.">pow</a> ( <span class="keywordtype">double</span> p ) {it_error ( <span class="stringliteral">"Not implemented"</span> );}; 
    101 <a name="l00065"></a>00065         }; 
    102 <a name="l00066"></a>00066  
    103 <a name="l00073"></a><a class="code" href="classbdm_1_1mEF.html">00073</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1mEF.html" title="Exponential family model.">mEF</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mpdf.html" title="Conditional probability density, e.g. modeling some dependencies.">mpdf</a> 
    104 <a name="l00074"></a>00074         { 
    105 <a name="l00075"></a>00075  
    106 <a name="l00076"></a>00076                 <span class="keyword">public</span>: 
    107 <a name="l00078"></a><a class="code" href="classbdm_1_1mEF.html#dd7caad7026b778af66e34480eec6f9e">00078</a>                         <a class="code" href="classbdm_1_1mEF.html#dd7caad7026b778af66e34480eec6f9e" title="Default constructor.">mEF</a> ( ) :<a class="code" href="classbdm_1_1mpdf.html" title="Conditional probability density, e.g. modeling some dependencies.">mpdf</a> ( ) {}; 
    108 <a name="l00079"></a>00079         }; 
    109 <a name="l00080"></a>00080  
    110 <a name="l00082"></a><a class="code" href="classbdm_1_1BMEF.html">00082</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1BM.html" title="Bayesian Model of a system, i.e. all uncertainty is modeled by probabilities.">BM</a> 
    111 <a name="l00083"></a>00083         { 
    112 <a name="l00084"></a>00084                 <span class="keyword">protected</span>: 
    113 <a name="l00086"></a><a class="code" href="classbdm_1_1BMEF.html#1331865e10fb1ccef65bb4c47fa3be64">00086</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1BMEF.html#1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a>; 
    114 <a name="l00088"></a><a class="code" href="classbdm_1_1BMEF.html#06e7b3ac03e10017d4288c76888e2865">00088</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1BMEF.html#06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a>; 
    115 <a name="l00089"></a>00089                 <span class="keyword">public</span>: 
    116 <a name="l00091"></a><a class="code" href="classbdm_1_1BMEF.html#2def512872ed8a4fc3b702371ec0be55">00091</a>                         <a class="code" href="classbdm_1_1BMEF.html#2def512872ed8a4fc3b702371ec0be55" title="Default constructor (=empty constructor).">BMEF</a> ( <span class="keywordtype">double</span> frg0=1.0 ) :<a class="code" href="classbdm_1_1BM.html" title="Bayesian Model of a system, i.e. all uncertainty is modeled by probabilities.">BM</a> ( ), <a class="code" href="classbdm_1_1BMEF.html#1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a> ( frg0 ) {} 
    117 <a name="l00093"></a><a class="code" href="classbdm_1_1BMEF.html#9662379513101405e159e76717104e62">00093</a>                         <a class="code" href="classbdm_1_1BMEF.html#9662379513101405e159e76717104e62" title="Copy constructor.">BMEF</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a> &amp;B ) :<a class="code" href="classbdm_1_1BM.html" title="Bayesian Model of a system, i.e. all uncertainty is modeled by probabilities.">BM</a> ( B ), <a class="code" href="classbdm_1_1BMEF.html#1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a> ( B.<a class="code" href="classbdm_1_1BMEF.html#1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a> ), <a class="code" href="classbdm_1_1BMEF.html#06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a> ( B.<a class="code" href="classbdm_1_1BMEF.html#06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a> ) {} 
    118 <a name="l00095"></a><a class="code" href="classbdm_1_1BMEF.html#d2b528b7a41ca67163152142f5404051">00095</a>                         <span class="keyword">virtual</span> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1BMEF.html#d2b528b7a41ca67163152142f5404051" title="get statistics from another model">set_statistics</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a>* BM0 ) {it_error ( <span class="stringliteral">"Not implemented"</span> );}; 
    119 <a name="l00097"></a><a class="code" href="classbdm_1_1BMEF.html#bf58deb99af2a6cc674f13ff90300de6">00097</a>                         <span class="keyword">virtual</span> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1BMEF.html#bf58deb99af2a6cc674f13ff90300de6" title="Weighted update of sufficient statistics (Bayes rule).">bayes</a> ( <span class="keyword">const</span> vec &amp;data, <span class="keyword">const</span> <span class="keywordtype">double</span> w ) {}; 
    120 <a name="l00098"></a>00098                         <span class="comment">//original Bayes</span> 
    121 <a name="l00099"></a>00099                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1BMEF.html#bf58deb99af2a6cc674f13ff90300de6" title="Weighted update of sufficient statistics (Bayes rule).">bayes</a> ( <span class="keyword">const</span> vec &amp;dt ); 
    122 <a name="l00101"></a><a class="code" href="classbdm_1_1BMEF.html#b2916a2e71a958665054473124d5e749">00101</a>                         <span class="keyword">virtual</span> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1BMEF.html#b2916a2e71a958665054473124d5e749" title="Flatten the posterior according to the given BMEF (of the same type!).">flatten</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a> * B ) {it_error ( <span class="stringliteral">"Not implemented"</span> );} 
    123 <a name="l00103"></a>00103 <span class="comment">//      virtual void flatten ( double nu0 ) {it_error ( "Not implemented" );}</span> 
    124 <a name="l00104"></a>00104  
    125 <a name="l00105"></a><a class="code" href="classbdm_1_1BMEF.html#62d2e4691bed41a1efa6b9c2e35e5c67">00105</a>                         <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a>* <a class="code" href="classbdm_1_1BMEF.html#62d2e4691bed41a1efa6b9c2e35e5c67" title="Flatten the posterior as if to keep nu0 data.">_copy_</a> ()<span class="keyword"> const </span>{it_error ( <span class="stringliteral">"function _copy_ not implemented for this BM"</span> ); <span class="keywordflow">return</span> NULL;}; 
    126 <a name="l00106"></a>00106         }; 
    127 <a name="l00107"></a>00107  
    128 <a name="l00108"></a>00108         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    129 <a name="l00109"></a>00109         <span class="keyword">class </span>mlnorm; 
    130 <a name="l00110"></a>00110  
    131 <a name="l00116"></a>00116         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    132 <a name="l00117"></a><a class="code" href="classbdm_1_1enorm.html">00117</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> 
    133 <a name="l00118"></a>00118         { 
    134 <a name="l00119"></a>00119                 <span class="keyword">protected</span>: 
    135 <a name="l00121"></a><a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7">00121</a>                         vec <a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a>; 
    136 <a name="l00123"></a><a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2">00123</a>                         sq_T <a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>; 
    137 <a name="l00124"></a>00124                 <span class="keyword">public</span>: 
    138 <a name="l00127"></a>00127  
    139 <a name="l00128"></a>00128                         <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm</a> ( ) :<a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> ( ), <a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a> ( ),<a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a> ( ) {}; 
    140 <a name="l00129"></a>00129                         <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm</a> ( <span class="keyword">const</span> vec &amp;<a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a>,<span class="keyword">const</span> sq_T &amp;<a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a> ) {set_parameters ( mu,R );} 
    141 <a name="l00130"></a>00130                         <span class="keywordtype">void</span> set_parameters ( <span class="keyword">const</span> vec &amp;<a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a>,<span class="keyword">const</span> sq_T &amp;<a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a> ); 
    142 <a name="l00131"></a>00131                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1enorm.html#61bd470764020bea6e1ed35000f259e6">from_setting</a>(<span class="keyword">const</span> Setting &amp;root); 
    143 <a name="l00132"></a><a class="code" href="classbdm_1_1enorm.html#38eb17ecb75d94a50b6782fcf735cfea">00132</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1enorm.html#38eb17ecb75d94a50b6782fcf735cfea" title="This method TODO.">validate</a>() { 
    144 <a name="l00133"></a>00133                                 it_assert(<a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a>.length()==<a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>.rows(),<span class="stringliteral">"parameters mismatch"</span>); 
    145 <a name="l00134"></a>00134                                 <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> = <a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a>.length(); 
    146 <a name="l00135"></a>00135                         } 
    147 <a name="l00137"></a>00137  
    148 <a name="l00140"></a>00140  
    149 <a name="l00142"></a>00142                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1enorm.html#d2e0d3a1e30ab3ab04df2d0c43ae74a2" title="dupdate in exponential form (not really handy)">dupdate</a> ( mat &amp;v,<span class="keywordtype">double</span> nu=1.0 ); 
    150 <a name="l00143"></a>00143  
    151 <a name="l00144"></a>00144                         vec <a class="code" href="classbdm_1_1enorm.html#e1a48f52351ec3a349bd443b713b1766" title="Returns a sample,  from density .">sample</a>() <span class="keyword">const</span>; 
     88<a name="l00043"></a><a class="code" href="classbdm_1_1eEF.html#ad5459d472d0feca7cf1fb5f65c4b9ef4">00043</a> <span class="comment"></span>                <a class="code" href="classbdm_1_1eEF.html#ad5459d472d0feca7cf1fb5f65c4b9ef4" title="default constructor">eEF</a> () : <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> () {}; 
     89<a name="l00045"></a>00045                 <span class="keyword">virtual</span> <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eEF.html#acd678fc9b02007a4b8d6692e746f1bf8" title="logarithm of the normalizing constant, ">lognc</a>() <span class="keyword">const</span> = 0; 
     90<a name="l00046"></a>00046  
     91<a name="l00048"></a><a class="code" href="classbdm_1_1eEF.html#aa4135778ecd9ab774762936c82a097c6">00048</a>                 <span class="keyword">virtual</span> <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eEF.html#aa4135778ecd9ab774762936c82a097c6" title="Evaluate normalized log-probability.">evallog_nn</a> (<span class="keyword">const</span> vec &amp;val)<span class="keyword"> const </span>{ 
     92<a name="l00049"></a>00049                         <a class="code" href="bdmerror_8h.html#a7c43f3a72afe68ab0c85663a1bb3521a" title="Unconditionally throw std::runtime_error.">bdm_error</a> (<span class="stringliteral">&quot;Not implemented&quot;</span>); 
     93<a name="l00050"></a>00050                         <span class="keywordflow">return</span> 0.0; 
     94<a name="l00051"></a>00051                 } 
     95<a name="l00052"></a>00052  
     96<a name="l00054"></a><a class="code" href="classbdm_1_1eEF.html#aa36d06ecdd6f4c79dc122510eaccc692">00054</a>                 <span class="keyword">virtual</span> <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eEF.html#aa36d06ecdd6f4c79dc122510eaccc692" title="Evaluate normalized log-probability.">evallog</a> (<span class="keyword">const</span> vec &amp;val)<span class="keyword"> const </span>{ 
     97<a name="l00055"></a>00055                         <span class="keywordtype">double</span> tmp; 
     98<a name="l00056"></a>00056                         tmp = <a class="code" href="classbdm_1_1eEF.html#aa4135778ecd9ab774762936c82a097c6" title="Evaluate normalized log-probability.">evallog_nn</a> (val) - <a class="code" href="classbdm_1_1eEF.html#acd678fc9b02007a4b8d6692e746f1bf8" title="logarithm of the normalizing constant, ">lognc</a>(); 
     99<a name="l00057"></a>00057                         <span class="keywordflow">return</span> tmp; 
     100<a name="l00058"></a>00058                 } 
     101<a name="l00060"></a><a class="code" href="classbdm_1_1eEF.html#a6886c60b6b690e503913240db5de0c6f">00060</a>                 <span class="keyword">virtual</span> vec <a class="code" href="classbdm_1_1eEF.html#a6886c60b6b690e503913240db5de0c6f" title="Evaluate normalized log-probability for many samples.">evallog_m</a> (<span class="keyword">const</span> mat &amp;Val)<span class="keyword"> const </span>{ 
     102<a name="l00061"></a>00061                         vec x (Val.cols()); 
     103<a name="l00062"></a>00062                         <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = 0;i &lt; Val.cols();i++) {x (i) = <a class="code" href="classbdm_1_1eEF.html#aa4135778ecd9ab774762936c82a097c6" title="Evaluate normalized log-probability.">evallog_nn</a> (Val.get_col (i)) ;} 
     104<a name="l00063"></a>00063                         <span class="keywordflow">return</span> x -<a class="code" href="classbdm_1_1eEF.html#acd678fc9b02007a4b8d6692e746f1bf8" title="logarithm of the normalizing constant, ">lognc</a>(); 
     105<a name="l00064"></a>00064                 } 
     106<a name="l00066"></a><a class="code" href="classbdm_1_1eEF.html#ad793f4fd6d0dcec5f16bff0ae45fc7d5">00066</a>                 <span class="keyword">virtual</span> vec <a class="code" href="classbdm_1_1eEF.html#ad793f4fd6d0dcec5f16bff0ae45fc7d5" title="Evaluate normalized log-probability for many samples.">evallog_m</a> (<span class="keyword">const</span> Array&lt;vec&gt; &amp;Val)<span class="keyword"> const </span>{ 
     107<a name="l00067"></a>00067                         vec x (Val.length()); 
     108<a name="l00068"></a>00068                         <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = 0;i &lt; Val.length();i++) {x (i) = <a class="code" href="classbdm_1_1eEF.html#aa4135778ecd9ab774762936c82a097c6" title="Evaluate normalized log-probability.">evallog_nn</a> (Val (i)) ;} 
     109<a name="l00069"></a>00069                         <span class="keywordflow">return</span> x -<a class="code" href="classbdm_1_1eEF.html#acd678fc9b02007a4b8d6692e746f1bf8" title="logarithm of the normalizing constant, ">lognc</a>(); 
     110<a name="l00070"></a>00070                 } 
     111<a name="l00071"></a>00071  
     112<a name="l00073"></a><a class="code" href="classbdm_1_1eEF.html#acf38af29e8e3d650c640509a52396053">00073</a>                 <span class="keyword">virtual</span> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eEF.html#acf38af29e8e3d650c640509a52396053" title="Power of the density, used e.g. to flatten the density.">pow</a> (<span class="keywordtype">double</span> p) { 
     113<a name="l00074"></a>00074                         <a class="code" href="bdmerror_8h.html#a7c43f3a72afe68ab0c85663a1bb3521a" title="Unconditionally throw std::runtime_error.">bdm_error</a> (<span class="stringliteral">&quot;Not implemented&quot;</span>); 
     114<a name="l00075"></a>00075                 } 
     115<a name="l00076"></a>00076 }; 
     116<a name="l00077"></a>00077  
     117<a name="l00078"></a>00078  
     118<a name="l00080"></a><a class="code" href="classbdm_1_1BMEF.html">00080</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1BM.html" title="Bayesian Model of a system, i.e. all uncertainty is modeled by probabilities.">BM</a> 
     119<a name="l00081"></a>00081 { 
     120<a name="l00082"></a>00082         <span class="keyword">protected</span>: 
     121<a name="l00084"></a><a class="code" href="classbdm_1_1BMEF.html#a1331865e10fb1ccef65bb4c47fa3be64">00084</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1BMEF.html#a1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a>; 
     122<a name="l00086"></a><a class="code" href="classbdm_1_1BMEF.html#a06e7b3ac03e10017d4288c76888e2865">00086</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1BMEF.html#a06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a>; 
     123<a name="l00087"></a>00087         <span class="keyword">public</span>: 
     124<a name="l00089"></a><a class="code" href="classbdm_1_1BMEF.html#a2def512872ed8a4fc3b702371ec0be55">00089</a>                 <a class="code" href="classbdm_1_1BMEF.html#a2def512872ed8a4fc3b702371ec0be55" title="Default constructor (=empty constructor).">BMEF</a> (<span class="keywordtype">double</span> frg0 = 1.0) : <a class="code" href="classbdm_1_1BM.html" title="Bayesian Model of a system, i.e. all uncertainty is modeled by probabilities.">BM</a> (), <a class="code" href="classbdm_1_1BMEF.html#a1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a> (frg0) {} 
     125<a name="l00091"></a><a class="code" href="classbdm_1_1BMEF.html#a9662379513101405e159e76717104e62">00091</a>                 <a class="code" href="classbdm_1_1BMEF.html#a9662379513101405e159e76717104e62" title="Copy constructor.">BMEF</a> (<span class="keyword">const</span> <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a> &amp;B) : <a class="code" href="classbdm_1_1BM.html" title="Bayesian Model of a system, i.e. all uncertainty is modeled by probabilities.">BM</a> (B), <a class="code" href="classbdm_1_1BMEF.html#a1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a> (B.<a class="code" href="classbdm_1_1BMEF.html#a1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a>), <a class="code" href="classbdm_1_1BMEF.html#a06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a> (B.<a class="code" href="classbdm_1_1BMEF.html#a06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a>) {} 
     126<a name="l00093"></a><a class="code" href="classbdm_1_1BMEF.html#ad2b528b7a41ca67163152142f5404051">00093</a>                 <span class="keyword">virtual</span> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1BMEF.html#ad2b528b7a41ca67163152142f5404051" title="get statistics from another model">set_statistics</a> (<span class="keyword">const</span> <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a>* BM0) { 
     127<a name="l00094"></a>00094                         <a class="code" href="bdmerror_8h.html#a7c43f3a72afe68ab0c85663a1bb3521a" title="Unconditionally throw std::runtime_error.">bdm_error</a> (<span class="stringliteral">&quot;Not implemented&quot;</span>); 
     128<a name="l00095"></a>00095                 } 
     129<a name="l00096"></a>00096  
     130<a name="l00098"></a><a class="code" href="classbdm_1_1BMEF.html#abf58deb99af2a6cc674f13ff90300de6">00098</a>                 <span class="keyword">virtual</span> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1BMEF.html#abf58deb99af2a6cc674f13ff90300de6" title="Weighted update of sufficient statistics (Bayes rule).">bayes</a> (<span class="keyword">const</span> vec &amp;data, <span class="keyword">const</span> <span class="keywordtype">double</span> w) {}; 
     131<a name="l00099"></a>00099                 <span class="comment">//original Bayes</span> 
     132<a name="l00100"></a>00100                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1BMEF.html#abf58deb99af2a6cc674f13ff90300de6" title="Weighted update of sufficient statistics (Bayes rule).">bayes</a> (<span class="keyword">const</span> vec &amp;dt); 
     133<a name="l00101"></a>00101  
     134<a name="l00103"></a><a class="code" href="classbdm_1_1BMEF.html#ab2916a2e71a958665054473124d5e749">00103</a>                 <span class="keyword">virtual</span> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1BMEF.html#ab2916a2e71a958665054473124d5e749" title="Flatten the posterior according to the given BMEF (of the same type!).">flatten</a> (<span class="keyword">const</span> <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a> * B) { 
     135<a name="l00104"></a>00104                         <a class="code" href="bdmerror_8h.html#a7c43f3a72afe68ab0c85663a1bb3521a" title="Unconditionally throw std::runtime_error.">bdm_error</a> (<span class="stringliteral">&quot;Not implemented&quot;</span>); 
     136<a name="l00105"></a>00105                 } 
     137<a name="l00106"></a>00106  
     138<a name="l00107"></a><a class="code" href="classbdm_1_1BMEF.html#a62d2e4691bed41a1efa6b9c2e35e5c67">00107</a>                 <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a>* <a class="code" href="classbdm_1_1BMEF.html#a62d2e4691bed41a1efa6b9c2e35e5c67">_copy_</a> ()<span class="keyword"> const </span>{ 
     139<a name="l00108"></a>00108                         <a class="code" href="bdmerror_8h.html#a7c43f3a72afe68ab0c85663a1bb3521a" title="Unconditionally throw std::runtime_error.">bdm_error</a> (<span class="stringliteral">&quot;function _copy_ not implemented for this BM&quot;</span>); 
     140<a name="l00109"></a>00109                         <span class="keywordflow">return</span> NULL; 
     141<a name="l00110"></a>00110                 } 
     142<a name="l00111"></a>00111 }; 
     143<a name="l00112"></a>00112  
     144<a name="l00113"></a>00113 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T, <span class="keyword">template</span> &lt;<span class="keyword">typename</span>&gt; <span class="keyword">class </span>TEpdf&gt; 
     145<a name="l00114"></a>00114 <span class="keyword">class </span>mlnorm; 
     146<a name="l00115"></a>00115  
     147<a name="l00121"></a>00121 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     148<a name="l00122"></a><a class="code" href="classbdm_1_1enorm.html">00122</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> 
     149<a name="l00123"></a>00123 { 
     150<a name="l00124"></a>00124         <span class="keyword">protected</span>: 
     151<a name="l00126"></a><a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7">00126</a>                 vec <a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a>; 
     152<a name="l00128"></a><a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2">00128</a>                 sq_T <a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>; 
     153<a name="l00129"></a>00129         <span class="keyword">public</span>: 
     154<a name="l00132"></a>00132  
     155<a name="l00133"></a>00133                 <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm</a> () : <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> (), <a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a> (), <a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a> () {}; 
     156<a name="l00134"></a>00134                 <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm</a> (<span class="keyword">const</span> vec &amp;<a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a>, <span class="keyword">const</span> sq_T &amp;<a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>) {set_parameters (mu, R);} 
     157<a name="l00135"></a>00135                 <span class="keywordtype">void</span> set_parameters (<span class="keyword">const</span> vec &amp;<a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a>, <span class="keyword">const</span> sq_T &amp;<a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>); 
     158<a name="l00136"></a>00136                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1enorm.html#a61bd470764020bea6e1ed35000f259e6">from_setting</a> (<span class="keyword">const</span> Setting &amp;root); 
     159<a name="l00137"></a><a class="code" href="classbdm_1_1enorm.html#a38eb17ecb75d94a50b6782fcf735cfea">00137</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1enorm.html#a38eb17ecb75d94a50b6782fcf735cfea" title="This method TODO.">validate</a>() { 
     160<a name="l00138"></a>00138                         <a class="code" href="bdmerror_8h.html#a89a0f906b242b79c5d3d342291b2cab4" title="Throw std::runtime_exception if t is not true and NDEBUG is not defined.">bdm_assert_debug</a> (<a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a>.length() == <a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>.rows(), <span class="stringliteral">&quot;parameters mismatch&quot;</span>); 
     161<a name="l00139"></a>00139                         <a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> = <a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a>.length(); 
     162<a name="l00140"></a>00140                 } 
     163<a name="l00142"></a>00142  
    152164<a name="l00145"></a>00145  
    153 <a name="l00146"></a>00146                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1enorm.html#e13aeed5b543b2179bacdc4fa2ae47a3" title="Evaluate normalized log-probability.">evallog_nn</a> ( <span class="keyword">const</span> vec &amp;val ) <span class="keyword">const</span>; 
    154 <a name="l00147"></a>00147                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1enorm.html#25785343aff102cc5df1cab08ba16d32" title="logarithm of the normalizing constant, ">lognc</a> () <span class="keyword">const</span>; 
    155 <a name="l00148"></a><a class="code" href="classbdm_1_1enorm.html#b2fa2915c35366392fe9bb022ca1a600">00148</a>                         vec <a class="code" href="classbdm_1_1enorm.html#b2fa2915c35366392fe9bb022ca1a600" title="return expected value">mean</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a>;} 
    156 <a name="l00149"></a><a class="code" href="classbdm_1_1enorm.html#729c75ef0fa8abae03d58ad1f81e6773">00149</a>                         vec <a class="code" href="classbdm_1_1enorm.html#729c75ef0fa8abae03d58ad1f81e6773" title="return expected variance (not covariance!)">variance</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> diag ( <a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>.to_mat() );} 
    157 <a name="l00150"></a>00150 <span class="comment">//      mlnorm&lt;sq_T&gt;* condition ( const RV &amp;rvn ) const ; &lt;=========== fails to cmpile. Why?</span> 
    158 <a name="l00151"></a>00151                         <a class="code" href="classbdm_1_1mpdf.html" title="Conditional probability density, e.g. modeling some dependencies.">mpdf</a>* <a class="code" href="classbdm_1_1enorm.html#baea4d49c657342b58297d68cda16d26" title="Return conditional density on the given RV, the remaining rvs will be in conditioning...">condition</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a> &amp;rvn ) <span class="keyword">const</span> ; 
    159 <a name="l00152"></a>00152                         <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm&lt;sq_T&gt;</a>* <a class="code" href="classbdm_1_1enorm.html#c2996bdaffad38fbe0fc776db54c9d80" title="Return marginal density on the given RV, the remainig rvs are intergrated out.">marginal</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a> &amp;<a class="code" href="classbdm_1_1epdf.html#62c5b8ff71d9ebe6cd58d3c342eb1dc8" title="Description of the random variable.">rv</a> ) <span class="keyword">const</span>; 
    160 <a name="l00153"></a>00153 <span class="comment">//                      epdf* marginal ( const RV &amp;rv ) const;</span> 
    161 <a name="l00155"></a>00155 <span class="comment"></span> 
    162 <a name="l00158"></a>00158  
    163 <a name="l00159"></a>00159                         vec&amp; _mu() {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a>;} 
    164 <a name="l00160"></a>00160                         <span class="keywordtype">void</span> set_mu ( <span class="keyword">const</span> vec mu0 ) { <a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a>=mu0;} 
    165 <a name="l00161"></a>00161                         sq_T&amp; _R() {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>;} 
    166 <a name="l00162"></a>00162                         <span class="keyword">const</span> sq_T&amp; _R()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>;} 
    167 <a name="l00164"></a>00164  
    168 <a name="l00165"></a>00165         }; 
    169 <a name="l00166"></a>00166         <a class="code" href="user__info_8h.html#4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a>(enorm&lt;chmat&gt;); 
    170 <a name="l00167"></a>00167         <a class="code" href="user__info_8h.html#4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a>(enorm&lt;ldmat&gt;); 
    171 <a name="l00168"></a>00168         <a class="code" href="user__info_8h.html#4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a>(enorm&lt;fsqmat&gt;); 
    172 <a name="l00169"></a>00169  
     165<a name="l00147"></a>00147                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1enorm.html#ad2e0d3a1e30ab3ab04df2d0c43ae74a2" title="dupdate in exponential form (not really handy)">dupdate</a> (mat &amp;v, <span class="keywordtype">double</span> nu = 1.0); 
     166<a name="l00148"></a>00148  
     167<a name="l00149"></a>00149                 vec <a class="code" href="classbdm_1_1enorm.html#ae1a48f52351ec3a349bd443b713b1766" title="Returns a sample,  from density .">sample</a>() <span class="keyword">const</span>; 
     168<a name="l00150"></a>00150  
     169<a name="l00151"></a>00151                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1enorm.html#ae13aeed5b543b2179bacdc4fa2ae47a3" title="Evaluate normalized log-probability.">evallog_nn</a> (<span class="keyword">const</span> vec &amp;val) <span class="keyword">const</span>; 
     170<a name="l00152"></a>00152                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1enorm.html#a25785343aff102cc5df1cab08ba16d32" title="logarithm of the normalizing constant, ">lognc</a> () <span class="keyword">const</span>; 
     171<a name="l00153"></a><a class="code" href="classbdm_1_1enorm.html#ab2fa2915c35366392fe9bb022ca1a600">00153</a>                 vec <a class="code" href="classbdm_1_1enorm.html#ab2fa2915c35366392fe9bb022ca1a600" title="return expected value">mean</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a>;} 
     172<a name="l00154"></a><a class="code" href="classbdm_1_1enorm.html#a729c75ef0fa8abae03d58ad1f81e6773">00154</a>                 vec <a class="code" href="classbdm_1_1enorm.html#a729c75ef0fa8abae03d58ad1f81e6773" title="return expected variance (not covariance!)">variance</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> diag (<a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>.to_mat());} 
     173<a name="l00155"></a>00155 <span class="comment">//      mlnorm&lt;sq_T&gt;* condition ( const RV &amp;rvn ) const ; &lt;=========== fails to cmpile. Why?</span> 
     174<a name="l00156"></a>00156                 <a class="code" href="classbdm_1_1shared__ptr.html">shared_ptr&lt;mpdf&gt;</a> <a class="code" href="classbdm_1_1enorm.html#a40201ca12c6900f7aeb493bb7c582e20" title="Return conditional density on the given RV, the remaining rvs will be in conditioning...">condition</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a> &amp;rvn ) <span class="keyword">const</span>; 
     175<a name="l00157"></a>00157  
     176<a name="l00158"></a>00158                 <span class="comment">// target not typed to mlnorm&lt;sq_T, enorm&lt;sq_T&gt; &gt; &amp;</span> 
     177<a name="l00159"></a>00159                 <span class="comment">// because that doesn&apos;t compile (perhaps because we</span> 
     178<a name="l00160"></a>00160                 <span class="comment">// haven&apos;t finished defining enorm yet), but the type</span> 
     179<a name="l00161"></a>00161                 <span class="comment">// is required</span> 
     180<a name="l00162"></a>00162                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1enorm.html#a40201ca12c6900f7aeb493bb7c582e20" title="Return conditional density on the given RV, the remaining rvs will be in conditioning...">condition</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a> &amp;rvn, <a class="code" href="classbdm_1_1mpdf.html" title="Conditional probability density, e.g. modeling , where  is random variable, rv, and...">mpdf</a> &amp;target ) <span class="keyword">const</span>; 
     181<a name="l00163"></a>00163  
     182<a name="l00164"></a>00164                 <a class="code" href="classbdm_1_1shared__ptr.html">shared_ptr&lt;epdf&gt;</a> <a class="code" href="classbdm_1_1enorm.html#adcb739443669ba0a8faf634196284c08" title="Return marginal density on the given RV, the remainig rvs are intergrated out.">marginal</a> (<span class="keyword">const</span> <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a> &amp;rvn ) <span class="keyword">const</span>; 
     183<a name="l00165"></a>00165                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1enorm.html#adcb739443669ba0a8faf634196284c08" title="Return marginal density on the given RV, the remainig rvs are intergrated out.">marginal</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a> &amp;rvn, <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm&lt;sq_T&gt;</a> &amp;target ) <span class="keyword">const</span>; 
     184<a name="l00167"></a>00167  
    173185<a name="l00170"></a>00170  
    174 <a name="l00177"></a><a class="code" href="classbdm_1_1egiw.html">00177</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1egiw.html" title="Gauss-inverse-Wishart density stored in LD form.">egiw</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> 
    175 <a name="l00178"></a>00178         { 
    176 <a name="l00179"></a>00179                 <span class="keyword">protected</span>: 
    177 <a name="l00181"></a><a class="code" href="classbdm_1_1egiw.html#ae56852845c6af176fd9017dbebbbd52">00181</a>                         <a class="code" href="classldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a> <a class="code" href="classbdm_1_1egiw.html#ae56852845c6af176fd9017dbebbbd52" title="Extended information matrix of sufficient statistics.">V</a>; 
    178 <a name="l00183"></a><a class="code" href="classbdm_1_1egiw.html#447eacf19d4f4083872686f044814dc4">00183</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1egiw.html#447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a>; 
    179 <a name="l00185"></a><a class="code" href="classbdm_1_1egiw.html#23e4d78bea7e98840f3da30e76a2b57a">00185</a>                         <span class="keywordtype">int</span> <a class="code" href="classbdm_1_1egiw.html#23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a>; 
    180 <a name="l00187"></a><a class="code" href="classbdm_1_1egiw.html#322414c32d9a21a006a5aab0311f64fd">00187</a>                         <span class="keywordtype">int</span> <a class="code" href="classbdm_1_1egiw.html#322414c32d9a21a006a5aab0311f64fd" title="Dimension of the regressor.">nPsi</a>; 
    181 <a name="l00188"></a>00188                 <span class="keyword">public</span>: 
    182 <a name="l00191"></a>00191                         <a class="code" href="classbdm_1_1egiw.html" title="Gauss-inverse-Wishart density stored in LD form.">egiw</a>() :<a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a>() {}; 
    183 <a name="l00192"></a>00192                         <a class="code" href="classbdm_1_1egiw.html" title="Gauss-inverse-Wishart density stored in LD form.">egiw</a> ( <span class="keywordtype">int</span> dimx0, <a class="code" href="classldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a> V0, <span class="keywordtype">double</span> nu0=-1.0 ) :<a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a>() {set_parameters ( dimx0,V0, nu0 );}; 
    184 <a name="l00193"></a>00193  
    185 <a name="l00194"></a>00194                         <span class="keywordtype">void</span> set_parameters ( <span class="keywordtype">int</span> dimx0, <a class="code" href="classldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a> V0, <span class="keywordtype">double</span> nu0=-1.0 ) 
    186 <a name="l00195"></a>00195                         { 
    187 <a name="l00196"></a>00196                                 <a class="code" href="classbdm_1_1egiw.html#23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a>=dimx0; 
    188 <a name="l00197"></a>00197                                 <a class="code" href="classbdm_1_1egiw.html#322414c32d9a21a006a5aab0311f64fd" title="Dimension of the regressor.">nPsi</a> = V0.<a class="code" href="classsqmat.html#071e80ced9cc3b8cbb360fa7462eb646" title="Reimplementing common functions of mat: rows().">rows</a>()-<a class="code" href="classbdm_1_1egiw.html#23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a>; 
    189 <a name="l00198"></a>00198                                 <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> = <a class="code" href="classbdm_1_1egiw.html#23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a>* ( <a class="code" href="classbdm_1_1egiw.html#23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a>+<a class="code" href="classbdm_1_1egiw.html#322414c32d9a21a006a5aab0311f64fd" title="Dimension of the regressor.">nPsi</a> ); <span class="comment">// size(R) + size(Theta)</span> 
    190 <a name="l00199"></a>00199  
    191 <a name="l00200"></a>00200                                 <a class="code" href="classbdm_1_1egiw.html#ae56852845c6af176fd9017dbebbbd52" title="Extended information matrix of sufficient statistics.">V</a>=V0; 
    192 <a name="l00201"></a>00201                                 <span class="keywordflow">if</span> ( nu0&lt;0 ) 
    193 <a name="l00202"></a>00202                                 { 
    194 <a name="l00203"></a>00203                                         <a class="code" href="classbdm_1_1egiw.html#447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a> = 0.1 +<a class="code" href="classbdm_1_1egiw.html#322414c32d9a21a006a5aab0311f64fd" title="Dimension of the regressor.">nPsi</a> +2*<a class="code" href="classbdm_1_1egiw.html#23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a> +2; <span class="comment">// +2 assures finite expected value of R</span> 
    195 <a name="l00204"></a>00204                                         <span class="comment">// terms before that are sufficient for finite normalization</span> 
    196 <a name="l00205"></a>00205                                 } 
    197 <a name="l00206"></a>00206                                 <span class="keywordflow">else</span> 
    198 <a name="l00207"></a>00207                                 { 
    199 <a name="l00208"></a>00208                                         <a class="code" href="classbdm_1_1egiw.html#447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a>=nu0; 
    200 <a name="l00209"></a>00209                                 } 
    201 <a name="l00210"></a>00210                         } 
    202 <a name="l00212"></a>00212  
    203 <a name="l00213"></a>00213                         vec <a class="code" href="classbdm_1_1egiw.html#920f21548b7a3723923dd108fe514c61" title="Returns a sample,  from density .">sample</a>() <span class="keyword">const</span>; 
    204 <a name="l00214"></a>00214                         vec <a class="code" href="classbdm_1_1egiw.html#df70c05f918c3a6f86d60f10c1fd6ba2" title="return expected value">mean</a>() <span class="keyword">const</span>; 
    205 <a name="l00215"></a>00215                         vec <a class="code" href="classbdm_1_1egiw.html#c1ecc406613cc2341225dc10c3d3b46a" title="return expected variance (not covariance!)">variance</a>() <span class="keyword">const</span>; 
    206 <a name="l00216"></a>00216  
    207 <a name="l00218"></a>00218                         vec <a class="code" href="classbdm_1_1egiw.html#66d2ba9295c306012b309efcc9e516f0" title="LS estimate of .">est_theta</a>() <span class="keyword">const</span>; 
    208 <a name="l00219"></a>00219  
    209 <a name="l00221"></a>00221                         <a class="code" href="classldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a> <a class="code" href="classbdm_1_1egiw.html#88c321a2051d1afdbb31a098896a717b" title="Covariance of the LS estimate.">est_theta_cov</a>() <span class="keyword">const</span>; 
    210 <a name="l00222"></a>00222  
    211 <a name="l00223"></a>00223                         <span class="keywordtype">void</span> mean_mat ( mat &amp;M, mat&amp;R ) <span class="keyword">const</span>; 
    212 <a name="l00225"></a>00225                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1egiw.html#bfb8e7c619b34ad804a73bff71742b5e" title="In this instance, val= [theta, r]. For multivariate instances, it is stored columnwise...">evallog_nn</a> ( <span class="keyword">const</span> vec &amp;val ) <span class="keyword">const</span>; 
    213 <a name="l00226"></a>00226                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1egiw.html#41d72ba7b2abc8a9a4209ffa98ed5633" title="logarithm of the normalizing constant, ">lognc</a> () <span class="keyword">const</span>; 
    214 <a name="l00227"></a><a class="code" href="classbdm_1_1egiw.html#8e610e95401a11baf34f65e16ecd87be">00227</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1egiw.html#8e610e95401a11baf34f65e16ecd87be" title="Power of the density, used e.g. to flatten the density.">pow</a> ( <span class="keywordtype">double</span> p ) {V*=p;<a class="code" href="classbdm_1_1egiw.html#447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a>*=p;}; 
     186<a name="l00171"></a>00171                 vec&amp; _mu() {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a>;} 
     187<a name="l00172"></a>00172                 <span class="keyword">const</span> vec&amp; _mu()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a>;} 
     188<a name="l00173"></a>00173                 <span class="keywordtype">void</span> set_mu (<span class="keyword">const</span> vec mu0) { <a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a> = mu0;} 
     189<a name="l00174"></a>00174                 sq_T&amp; _R() {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>;} 
     190<a name="l00175"></a>00175                 <span class="keyword">const</span> sq_T&amp; _R()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>;} 
     191<a name="l00177"></a>00177  
     192<a name="l00178"></a>00178 }; 
     193<a name="l00179"></a>00179 <a class="code" href="user__info_8h.html#ab06f1e7c44a8306e321f62a0099210b9" title="Variant of UIREGISTER for templated classes.">UIREGISTER2</a> (enorm, chmat); 
     194<a name="l00180"></a>00180 SHAREDPTR2 ( enorm, chmat ); 
     195<a name="l00181"></a>00181 <a class="code" href="user__info_8h.html#ab06f1e7c44a8306e321f62a0099210b9" title="Variant of UIREGISTER for templated classes.">UIREGISTER2</a> (enorm, ldmat); 
     196<a name="l00182"></a>00182 SHAREDPTR2 ( enorm, ldmat ); 
     197<a name="l00183"></a>00183 <a class="code" href="user__info_8h.html#ab06f1e7c44a8306e321f62a0099210b9" title="Variant of UIREGISTER for templated classes.">UIREGISTER2</a> (enorm, fsqmat); 
     198<a name="l00184"></a>00184 SHAREDPTR2 ( enorm, fsqmat ); 
     199<a name="l00185"></a>00185  
     200<a name="l00186"></a>00186  
     201<a name="l00193"></a><a class="code" href="classbdm_1_1egiw.html">00193</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1egiw.html" title="Gauss-inverse-Wishart density stored in LD form.">egiw</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> 
     202<a name="l00194"></a>00194 { 
     203<a name="l00195"></a>00195         <span class="keyword">protected</span>: 
     204<a name="l00197"></a><a class="code" href="classbdm_1_1egiw.html#aae56852845c6af176fd9017dbebbbd52">00197</a>                 <a class="code" href="classbdm_1_1ldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a> <a class="code" href="classbdm_1_1egiw.html#aae56852845c6af176fd9017dbebbbd52" title="Extended information matrix of sufficient statistics.">V</a>; 
     205<a name="l00199"></a><a class="code" href="classbdm_1_1egiw.html#a447eacf19d4f4083872686f044814dc4">00199</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1egiw.html#a447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a>; 
     206<a name="l00201"></a><a class="code" href="classbdm_1_1egiw.html#a23e4d78bea7e98840f3da30e76a2b57a">00201</a>                 <span class="keywordtype">int</span> <a class="code" href="classbdm_1_1egiw.html#a23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a>; 
     207<a name="l00203"></a><a class="code" href="classbdm_1_1egiw.html#a322414c32d9a21a006a5aab0311f64fd">00203</a>                 <span class="keywordtype">int</span> <a class="code" href="classbdm_1_1egiw.html#a322414c32d9a21a006a5aab0311f64fd" title="Dimension of the regressor.">nPsi</a>; 
     208<a name="l00204"></a>00204         <span class="keyword">public</span>: 
     209<a name="l00207"></a>00207                 <a class="code" href="classbdm_1_1egiw.html" title="Gauss-inverse-Wishart density stored in LD form.">egiw</a>() : <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a>() {}; 
     210<a name="l00208"></a>00208                 <a class="code" href="classbdm_1_1egiw.html" title="Gauss-inverse-Wishart density stored in LD form.">egiw</a> (<span class="keywordtype">int</span> dimx0, <a class="code" href="classbdm_1_1ldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a> V0, <span class="keywordtype">double</span> nu0 = -1.0) : <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a>() {set_parameters (dimx0, V0, nu0);}; 
     211<a name="l00209"></a>00209  
     212<a name="l00210"></a>00210                 <span class="keywordtype">void</span> set_parameters (<span class="keywordtype">int</span> dimx0, ldmat V0, <span class="keywordtype">double</span> nu0 = -1.0) { 
     213<a name="l00211"></a>00211                         <a class="code" href="classbdm_1_1egiw.html#a23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a> = dimx0; 
     214<a name="l00212"></a>00212                         <a class="code" href="classbdm_1_1egiw.html#a322414c32d9a21a006a5aab0311f64fd" title="Dimension of the regressor.">nPsi</a> = V0.rows() - <a class="code" href="classbdm_1_1egiw.html#a23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a>; 
     215<a name="l00213"></a>00213                         <a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> = <a class="code" href="classbdm_1_1egiw.html#a23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a> * (<a class="code" href="classbdm_1_1egiw.html#a23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a> + <a class="code" href="classbdm_1_1egiw.html#a322414c32d9a21a006a5aab0311f64fd" title="Dimension of the regressor.">nPsi</a>); <span class="comment">// size(R) + size(Theta)</span> 
     216<a name="l00214"></a>00214  
     217<a name="l00215"></a>00215                         <a class="code" href="classbdm_1_1egiw.html#aae56852845c6af176fd9017dbebbbd52" title="Extended information matrix of sufficient statistics.">V</a> = V0; 
     218<a name="l00216"></a>00216                         <span class="keywordflow">if</span> (nu0 &lt; 0) { 
     219<a name="l00217"></a>00217                                 <a class="code" href="classbdm_1_1egiw.html#a447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a> = 0.1 + <a class="code" href="classbdm_1_1egiw.html#a322414c32d9a21a006a5aab0311f64fd" title="Dimension of the regressor.">nPsi</a> + 2 * <a class="code" href="classbdm_1_1egiw.html#a23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a> + 2; <span class="comment">// +2 assures finite expected value of R</span> 
     220<a name="l00218"></a>00218                                 <span class="comment">// terms before that are sufficient for finite normalization</span> 
     221<a name="l00219"></a>00219                         } <span class="keywordflow">else</span> { 
     222<a name="l00220"></a>00220                                 <a class="code" href="classbdm_1_1egiw.html#a447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a> = nu0; 
     223<a name="l00221"></a>00221                         } 
     224<a name="l00222"></a>00222                 } 
     225<a name="l00224"></a>00224  
     226<a name="l00225"></a>00225                 vec <a class="code" href="classbdm_1_1egiw.html#a920f21548b7a3723923dd108fe514c61" title="Returns a sample,  from density .">sample</a>() <span class="keyword">const</span>; 
     227<a name="l00226"></a>00226                 vec <a class="code" href="classbdm_1_1egiw.html#adf70c05f918c3a6f86d60f10c1fd6ba2" title="return expected value">mean</a>() <span class="keyword">const</span>; 
     228<a name="l00227"></a>00227                 vec <a class="code" href="classbdm_1_1egiw.html#ac1ecc406613cc2341225dc10c3d3b46a" title="return expected variance (not covariance!)">variance</a>() <span class="keyword">const</span>; 
    215229<a name="l00228"></a>00228  
     230<a name="l00230"></a>00230                 vec <a class="code" href="classbdm_1_1egiw.html#a66d2ba9295c306012b309efcc9e516f0" title="LS estimate of .">est_theta</a>() <span class="keyword">const</span>; 
    216231<a name="l00231"></a>00231  
    217 <a name="l00232"></a>00232                         <a class="code" href="classldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a>&amp; _V() {<span class="keywordflow">return</span> V;} 
    218 <a name="l00233"></a>00233                         <span class="keyword">const</span> <a class="code" href="classldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a>&amp; _V()<span class="keyword"> const </span>{<span class="keywordflow">return</span> V;} 
    219 <a name="l00234"></a>00234                         <span class="keywordtype">double</span>&amp; _nu()  {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1egiw.html#447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a>;} 
    220 <a name="l00235"></a>00235                         <span class="keyword">const</span> <span class="keywordtype">double</span>&amp; _nu()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1egiw.html#447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a>;} 
    221 <a name="l00236"></a><a class="code" href="classbdm_1_1egiw.html#55b76ec75bd2df5ef9cab3be20a33bbb">00236</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1egiw.html#55b76ec75bd2df5ef9cab3be20a33bbb">from_setting</a>(<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>){ 
    222 <a name="l00237"></a>00237                                 <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>(<a class="code" href="classbdm_1_1egiw.html#447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a>,<span class="keyword">set</span>,<span class="stringliteral">"nu"</span>); 
    223 <a name="l00238"></a>00238                                 <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>(<a class="code" href="classbdm_1_1egiw.html#23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a>,<span class="keyword">set</span>,<span class="stringliteral">"dimx"</span>); 
    224 <a name="l00239"></a>00239                                 <span class="keyword">set</span>.lookupValue(<span class="stringliteral">"nu"</span>,<a class="code" href="classbdm_1_1egiw.html#447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a>); 
    225 <a name="l00240"></a>00240                                 <span class="keyword">set</span>.lookupValue(<span class="stringliteral">"dimx"</span>,<a class="code" href="classbdm_1_1egiw.html#23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a>); 
    226 <a name="l00241"></a>00241                                 mat V; 
    227 <a name="l00242"></a>00242                                 <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>(V,<span class="keyword">set</span>,<span class="stringliteral">"V"</span>, UI::compulsory); 
    228 <a name="l00243"></a>00243                                 set_parameters(<a class="code" href="classbdm_1_1egiw.html#23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a>, V, <a class="code" href="classbdm_1_1egiw.html#447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a>); 
    229 <a name="l00244"></a>00244                                 <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a>* <a class="code" href="classbdm_1_1epdf.html#62c5b8ff71d9ebe6cd58d3c342eb1dc8" title="Description of the random variable.">rv</a>=UI::build&lt;RV&gt;(<span class="keyword">set</span>,<span class="stringliteral">"rv"</span>, UI::compulsory); 
    230 <a name="l00245"></a>00245                                 <a class="code" href="classbdm_1_1epdf.html#f423e28448dbb69ef4905295ec8de8ff" title="Name its rv.">set_rv</a>(*rv); 
    231 <a name="l00246"></a>00246                                 <span class="keyword">delete</span> rv; 
    232 <a name="l00247"></a>00247                         } 
    233 <a name="l00249"></a>00249         }; 
    234 <a name="l00250"></a>00250         <a class="code" href="user__info_8h.html#4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a>(egiw); 
    235 <a name="l00251"></a>00251  
    236 <a name="l00260"></a><a class="code" href="classbdm_1_1eDirich.html">00260</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1eDirich.html" title="Dirichlet posterior density.">eDirich</a>: <span class="keyword">public</span> <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> 
    237 <a name="l00261"></a>00261         { 
    238 <a name="l00262"></a>00262                 <span class="keyword">protected</span>: 
    239 <a name="l00264"></a><a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2">00264</a>                         vec <a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>; 
    240 <a name="l00265"></a>00265                 <span class="keyword">public</span>: 
    241 <a name="l00268"></a>00268  
    242 <a name="l00269"></a>00269                         <a class="code" href="classbdm_1_1eDirich.html" title="Dirichlet posterior density.">eDirich</a> () : <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> ( ) {}; 
    243 <a name="l00270"></a>00270                         <a class="code" href="classbdm_1_1eDirich.html" title="Dirichlet posterior density.">eDirich</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1eDirich.html" title="Dirichlet posterior density.">eDirich</a> &amp;D0 ) : <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> () {set_parameters ( D0.<a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a> );}; 
    244 <a name="l00271"></a>00271                         eDirich ( <span class="keyword">const</span> vec &amp;beta0 ) {set_parameters ( beta0 );}; 
    245 <a name="l00272"></a>00272                         <span class="keywordtype">void</span> set_parameters ( <span class="keyword">const</span> vec &amp;beta0 ) 
    246 <a name="l00273"></a>00273                         { 
    247 <a name="l00274"></a>00274                                 <a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>= beta0; 
    248 <a name="l00275"></a>00275                                 <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> = <a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>.length(); 
    249 <a name="l00276"></a>00276                         } 
    250 <a name="l00278"></a>00278  
    251 <a name="l00279"></a><a class="code" href="classbdm_1_1eDirich.html#3290613d31d58daa8a45a54b003871fc">00279</a>                         vec <a class="code" href="classbdm_1_1eDirich.html#3290613d31d58daa8a45a54b003871fc" title="Returns a sample,  from density .">sample</a>()<span class="keyword"> const </span>{it_error ( <span class="stringliteral">"Not implemented"</span> );<span class="keywordflow">return</span> vec_1 ( 0.0 );}; 
    252 <a name="l00280"></a><a class="code" href="classbdm_1_1eDirich.html#cb343355ec791298bb5a3404cd482fb6">00280</a>                         vec <a class="code" href="classbdm_1_1eDirich.html#cb343355ec791298bb5a3404cd482fb6" title="return expected value">mean</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>/sum(<a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>);}; 
    253 <a name="l00281"></a><a class="code" href="classbdm_1_1eDirich.html#43c547a2507e233706f92712d8c2aacc">00281</a>                         vec <a class="code" href="classbdm_1_1eDirich.html#43c547a2507e233706f92712d8c2aacc" title="return expected variance (not covariance!)">variance</a>()<span class="keyword"> const </span>{<span class="keywordtype">double</span> gamma =sum(<a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>); <span class="keywordflow">return</span> elem_mult ( <a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>, ( <a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>+1 ) ) / ( gamma* ( gamma+1 ) );} 
    254 <a name="l00283"></a><a class="code" href="classbdm_1_1eDirich.html#e09a24938e80c3d94b0ee842d1552318">00283</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eDirich.html#e09a24938e80c3d94b0ee842d1552318" title="In this instance, val is ...">evallog_nn</a> ( <span class="keyword">const</span> vec &amp;val )<span class="keyword"> const</span> 
    255 <a name="l00284"></a>00284 <span class="keyword">                        </span>{ 
    256 <a name="l00285"></a>00285                                 <span class="keywordtype">double</span> tmp; tmp= ( <a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>-1 ) *log ( val ); 
    257 <a name="l00286"></a>00286 <span class="comment">//                              it_assert_debug ( std::isfinite ( tmp ),"Infinite value" );</span> 
    258 <a name="l00287"></a>00287                                 <span class="keywordflow">return</span> tmp; 
    259 <a name="l00288"></a>00288                         }; 
    260 <a name="l00289"></a><a class="code" href="classbdm_1_1eDirich.html#279a99f6266c82fe2273e83841f19eb2">00289</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eDirich.html#279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a> ()<span class="keyword"> const</span> 
    261 <a name="l00290"></a>00290 <span class="keyword">                        </span>{ 
    262 <a name="l00291"></a>00291                                 <span class="keywordtype">double</span> tmp; 
    263 <a name="l00292"></a>00292                                 <span class="keywordtype">double</span> gam=sum ( <a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a> ); 
    264 <a name="l00293"></a>00293                                 <span class="keywordtype">double</span> lgb=0.0; 
    265 <a name="l00294"></a>00294                                 <span class="keywordflow">for</span> ( <span class="keywordtype">int</span> i=0;i&lt;<a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>.length();i++ ) {lgb+=lgamma ( <a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a> ( i ) );} 
    266 <a name="l00295"></a>00295                                 tmp= lgb-lgamma ( gam ); 
    267 <a name="l00296"></a>00296 <span class="comment">//                              it_assert_debug ( std::isfinite ( tmp ),"Infinite value" );</span> 
    268 <a name="l00297"></a>00297                                 <span class="keywordflow">return</span> tmp; 
    269 <a name="l00298"></a>00298                         }; 
    270 <a name="l00300"></a><a class="code" href="classbdm_1_1eDirich.html#175e0add26d2105c28d8121eefb9e324">00300</a>                         vec&amp; <a class="code" href="classbdm_1_1eDirich.html#175e0add26d2105c28d8121eefb9e324" title="access function">_beta</a>()  {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1eDirich.html#f25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>;} 
    271 <a name="l00302"></a>00302         }; 
    272 <a name="l00303"></a>00303  
    273 <a name="l00305"></a><a class="code" href="classbdm_1_1multiBM.html">00305</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1multiBM.html" title="Estimator for Multinomial density.">multiBM</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a> 
    274 <a name="l00306"></a>00306         { 
    275 <a name="l00307"></a>00307                 <span class="keyword">protected</span>: 
    276 <a name="l00309"></a><a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a">00309</a>                         <a class="code" href="classbdm_1_1eDirich.html" title="Dirichlet posterior density.">eDirich</a> <a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>; 
    277 <a name="l00311"></a><a class="code" href="classbdm_1_1multiBM.html#044263356944c92209eecd39a5187d25">00311</a>                         vec &amp;<a class="code" href="classbdm_1_1multiBM.html#044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a>; 
    278 <a name="l00312"></a>00312                 <span class="keyword">public</span>: 
    279 <a name="l00314"></a><a class="code" href="classbdm_1_1multiBM.html#c4dd6d9522a8a605776d21bac9bd9daf">00314</a>                         <a class="code" href="classbdm_1_1multiBM.html#c4dd6d9522a8a605776d21bac9bd9daf" title="Default constructor.">multiBM</a> ( ) : <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a> ( ),<a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a> ( ),<a class="code" href="classbdm_1_1multiBM.html#044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a> ( <a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>._beta() ) 
    280 <a name="l00315"></a>00315                         { 
    281 <a name="l00316"></a>00316                                 <span class="keywordflow">if</span> ( <a class="code" href="classbdm_1_1multiBM.html#044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a>.length() &gt;0 ) {<a class="code" href="classbdm_1_1BMEF.html#06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a>=<a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>.<a class="code" href="classbdm_1_1eDirich.html#279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>();} 
    282 <a name="l00317"></a>00317                                 <span class="keywordflow">else</span>{<a class="code" href="classbdm_1_1BMEF.html#06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a>=0.0;} 
    283 <a name="l00318"></a>00318                         } 
    284 <a name="l00320"></a><a class="code" href="classbdm_1_1multiBM.html#c4378cf8037f6bed29c74eea63344b31">00320</a>                         <a class="code" href="classbdm_1_1multiBM.html#c4378cf8037f6bed29c74eea63344b31" title="Copy constructor.">multiBM</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1multiBM.html" title="Estimator for Multinomial density.">multiBM</a> &amp;B ) : <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a> ( B ),<a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a> ( B.<a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a> ),<a class="code" href="classbdm_1_1multiBM.html#044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a> ( <a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>._beta() ) {} 
    285 <a name="l00322"></a><a class="code" href="classbdm_1_1multiBM.html#dbe6b90d410dc062a233d1dc09eeba52">00322</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1multiBM.html#dbe6b90d410dc062a233d1dc09eeba52" title="Sets sufficient statistics to match that of givefrom mB0.">set_statistics</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1BM.html" title="Bayesian Model of a system, i.e. all uncertainty is modeled by probabilities.">BM</a>* mB0 ) {<span class="keyword">const</span> <a class="code" href="classbdm_1_1multiBM.html" title="Estimator for Multinomial density.">multiBM</a>* mB=<span class="keyword">dynamic_cast&lt;</span><span class="keyword">const </span><a class="code" href="classbdm_1_1multiBM.html" title="Estimator for Multinomial density.">multiBM</a>*<span class="keyword">&gt;</span> ( mB0 ); <a class="code" href="classbdm_1_1multiBM.html#044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a>=mB-&gt;<a class="code" href="classbdm_1_1multiBM.html#044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a>;} 
    286 <a name="l00323"></a><a class="code" href="classbdm_1_1multiBM.html#1e4bf41b61937fd80f34049742e23f95">00323</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1multiBM.html#1e4bf41b61937fd80f34049742e23f95" title="Incremental Bayes rule.">bayes</a> ( <span class="keyword">const</span> vec &amp;dt ) 
    287 <a name="l00324"></a>00324                         { 
    288 <a name="l00325"></a>00325                                 <span class="keywordflow">if</span> ( <a class="code" href="classbdm_1_1BMEF.html#1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a>&lt;1.0 ) {<a class="code" href="classbdm_1_1multiBM.html#044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a>*=<a class="code" href="classbdm_1_1BMEF.html#1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a>;<a class="code" href="classbdm_1_1BMEF.html#06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a>=<a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>.<a class="code" href="classbdm_1_1eDirich.html#279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>();} 
    289 <a name="l00326"></a>00326                                 <a class="code" href="classbdm_1_1multiBM.html#044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a>+=dt; 
    290 <a name="l00327"></a>00327                                 <span class="keywordflow">if</span> ( <a class="code" href="classbdm_1_1BM.html#faff0ad12556fe7dc0e2807d4fd938ee" title="If true, the filter will compute likelihood of the data record and store it in ll...">evalll</a> ) {<a class="code" href="classbdm_1_1BM.html#4064b6559d962633e4372b12f4cd204a" title="Logarithm of marginalized data likelihood.">ll</a>=<a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>.<a class="code" href="classbdm_1_1eDirich.html#279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>()-<a class="code" href="classbdm_1_1BMEF.html#06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a>;} 
    291 <a name="l00328"></a>00328                         } 
    292 <a name="l00329"></a><a class="code" href="classbdm_1_1multiBM.html#e157b607c1e3fa91d42aeea44458e2bf">00329</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1multiBM.html#e157b607c1e3fa91d42aeea44458e2bf">logpred</a> ( <span class="keyword">const</span> vec &amp;dt )<span class="keyword"> const</span> 
    293 <a name="l00330"></a>00330 <span class="keyword">                        </span>{ 
    294 <a name="l00331"></a>00331                                 <a class="code" href="classbdm_1_1eDirich.html" title="Dirichlet posterior density.">eDirich</a> pred ( <a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a> ); 
    295 <a name="l00332"></a>00332                                 vec &amp;<a class="code" href="classbdm_1_1multiBM.html#044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a> = pred.<a class="code" href="classbdm_1_1eDirich.html#175e0add26d2105c28d8121eefb9e324" title="access function">_beta</a>(); 
    296 <a name="l00333"></a>00333  
    297 <a name="l00334"></a>00334                                 <span class="keywordtype">double</span> lll; 
    298 <a name="l00335"></a>00335                                 <span class="keywordflow">if</span> ( <a class="code" href="classbdm_1_1BMEF.html#1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a>&lt;1.0 ) 
    299 <a name="l00336"></a>00336                                         {beta*=<a class="code" href="classbdm_1_1BMEF.html#1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a>;lll=pred.<a class="code" href="classbdm_1_1eDirich.html#279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>();} 
    300 <a name="l00337"></a>00337                                 <span class="keywordflow">else</span> 
    301 <a name="l00338"></a>00338                                         <span class="keywordflow">if</span> ( <a class="code" href="classbdm_1_1BM.html#faff0ad12556fe7dc0e2807d4fd938ee" title="If true, the filter will compute likelihood of the data record and store it in ll...">evalll</a> ) {lll=<a class="code" href="classbdm_1_1BMEF.html#06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a>;} 
    302 <a name="l00339"></a>00339                                         <span class="keywordflow">else</span>{lll=pred.<a class="code" href="classbdm_1_1eDirich.html#279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>();} 
    303 <a name="l00340"></a>00340  
    304 <a name="l00341"></a>00341                                 beta+=dt; 
    305 <a name="l00342"></a>00342                                 <span class="keywordflow">return</span> pred.<a class="code" href="classbdm_1_1eDirich.html#279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>()-lll; 
    306 <a name="l00343"></a>00343                         } 
    307 <a name="l00344"></a><a class="code" href="classbdm_1_1multiBM.html#aaeb18c989088feb8d26d300e4971732">00344</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1multiBM.html#aaeb18c989088feb8d26d300e4971732" title="Flatten the posterior according to the given BMEF (of the same type!).">flatten</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a>* B ) 
    308 <a name="l00345"></a>00345                         { 
    309 <a name="l00346"></a>00346                                 <span class="keyword">const</span> <a class="code" href="classbdm_1_1multiBM.html" title="Estimator for Multinomial density.">multiBM</a>* E=<span class="keyword">dynamic_cast&lt;</span><span class="keyword">const </span><a class="code" href="classbdm_1_1multiBM.html" title="Estimator for Multinomial density.">multiBM</a>*<span class="keyword">&gt;</span> ( B ); 
    310 <a name="l00347"></a>00347                                 <span class="comment">// sum(beta) should be equal to sum(B.beta)</span> 
    311 <a name="l00348"></a>00348                                 <span class="keyword">const</span> vec &amp;Eb=E-&gt;<a class="code" href="classbdm_1_1multiBM.html#044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a>;<span class="comment">//const_cast&lt;multiBM*&gt; ( E )-&gt;_beta();</span> 
    312 <a name="l00349"></a>00349                                 <a class="code" href="classbdm_1_1multiBM.html#044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a>*= ( sum ( Eb ) /sum ( <a class="code" href="classbdm_1_1multiBM.html#044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a> ) ); 
    313 <a name="l00350"></a>00350                                 <span class="keywordflow">if</span> ( <a class="code" href="classbdm_1_1BM.html#faff0ad12556fe7dc0e2807d4fd938ee" title="If true, the filter will compute likelihood of the data record and store it in ll...">evalll</a> ) {<a class="code" href="classbdm_1_1BMEF.html#06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a>=<a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>.<a class="code" href="classbdm_1_1eDirich.html#279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>();} 
    314 <a name="l00351"></a>00351                         } 
    315 <a name="l00352"></a>00352                         <span class="keyword">const</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a>&amp; posterior()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>;}; 
    316 <a name="l00353"></a>00353                         <span class="keyword">const</span> eDirich* _e()<span class="keyword"> const </span>{<span class="keywordflow">return</span> &amp;<a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>;}; 
    317 <a name="l00354"></a>00354                         <span class="keywordtype">void</span> set_parameters ( <span class="keyword">const</span> vec &amp;beta0 ) 
    318 <a name="l00355"></a>00355                         { 
    319 <a name="l00356"></a>00356                                 <a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>.set_parameters ( beta0 ); 
    320 <a name="l00357"></a>00357                                 <span class="keywordflow">if</span> ( <a class="code" href="classbdm_1_1BM.html#faff0ad12556fe7dc0e2807d4fd938ee" title="If true, the filter will compute likelihood of the data record and store it in ll...">evalll</a> ) {<a class="code" href="classbdm_1_1BMEF.html#06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a>=<a class="code" href="classbdm_1_1multiBM.html#9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>.lognc();} 
    321 <a name="l00358"></a>00358                         } 
    322 <a name="l00359"></a>00359         }; 
    323 <a name="l00360"></a>00360  
    324 <a name="l00370"></a><a class="code" href="classbdm_1_1egamma.html">00370</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1egamma.html" title="Gamma posterior density.">egamma</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> 
    325 <a name="l00371"></a>00371         { 
    326 <a name="l00372"></a>00372                 <span class="keyword">protected</span>: 
    327 <a name="l00374"></a><a class="code" href="classbdm_1_1egamma.html#0901ec983e66b8337aaa506e13b122fa">00374</a>                         vec <a class="code" href="classbdm_1_1egamma.html#0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>; 
    328 <a name="l00376"></a><a class="code" href="classbdm_1_1egamma.html#457bfb1ccb2057df85073e519a15ccc1">00376</a>                         vec <a class="code" href="classbdm_1_1egamma.html#457bfb1ccb2057df85073e519a15ccc1" title="Vector .">beta</a>; 
    329 <a name="l00377"></a>00377                 <span class="keyword">public</span> : 
    330 <a name="l00380"></a>00380                         <a class="code" href="classbdm_1_1egamma.html" title="Gamma posterior density.">egamma</a> ( ) :<a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> ( ), <a class="code" href="classbdm_1_1egamma.html#0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a> ( 0 ), <a class="code" href="classbdm_1_1egamma.html#457bfb1ccb2057df85073e519a15ccc1" title="Vector .">beta</a> ( 0 ) {}; 
    331 <a name="l00381"></a>00381                         <a class="code" href="classbdm_1_1egamma.html" title="Gamma posterior density.">egamma</a> ( <span class="keyword">const</span> vec &amp;a, <span class="keyword">const</span> vec &amp;b ) {set_parameters ( a, b );}; 
    332 <a name="l00382"></a>00382                         <span class="keywordtype">void</span> set_parameters ( <span class="keyword">const</span> vec &amp;a, <span class="keyword">const</span> vec &amp;b ) {<a class="code" href="classbdm_1_1egamma.html#0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>=a,<a class="code" href="classbdm_1_1egamma.html#457bfb1ccb2057df85073e519a15ccc1" title="Vector .">beta</a>=b;<a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> = <a class="code" href="classbdm_1_1egamma.html#0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>.length();}; 
    333 <a name="l00384"></a>00384  
    334 <a name="l00385"></a>00385                         vec <a class="code" href="classbdm_1_1egamma.html#6ed82f0fd05f6002487256d8e75a0bbd" title="Returns a sample,  from density .">sample</a>() <span class="keyword">const</span>; 
    335 <a name="l00387"></a>00387 <span class="comment">//      mat sample ( int N ) const;</span> 
    336 <a name="l00388"></a>00388                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1egamma.html#a8e11e5a580ff42a1b205974c60768c6" title="TODO: is it used anywhere?">evallog</a> ( <span class="keyword">const</span> vec &amp;val ) <span class="keyword">const</span>; 
    337 <a name="l00389"></a>00389                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1egamma.html#9a66cbd100e8520c769ccb3c451f86f8" title="logarithm of the normalizing constant, ">lognc</a> () <span class="keyword">const</span>; 
    338 <a name="l00391"></a><a class="code" href="classbdm_1_1egamma.html#0865cb3d6339fdc7410806cf70a329ed">00391</a>                         vec&amp; <a class="code" href="classbdm_1_1egamma.html#0865cb3d6339fdc7410806cf70a329ed" title="Returns poiter to alpha and beta. Potentially dengerous: use with care!">_alpha</a>() {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1egamma.html#0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>;} 
    339 <a name="l00392"></a>00392                         vec&amp; _beta() {<span class="keywordflow">return</span> beta;} 
    340 <a name="l00393"></a><a class="code" href="classbdm_1_1egamma.html#49d256c42cce14c6faa56ec242b57e85">00393</a>                         vec <a class="code" href="classbdm_1_1egamma.html#49d256c42cce14c6faa56ec242b57e85" title="return expected value">mean</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> elem_div ( <a class="code" href="classbdm_1_1egamma.html#0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>,beta );} 
    341 <a name="l00394"></a><a class="code" href="classbdm_1_1egamma.html#36986cc01917cd0796fadc17125bdec1">00394</a>                         vec <a class="code" href="classbdm_1_1egamma.html#36986cc01917cd0796fadc17125bdec1" title="return expected variance (not covariance!)">variance</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> elem_div ( <a class="code" href="classbdm_1_1egamma.html#0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>,elem_mult ( beta,beta ) ); } 
    342 <a name="l00395"></a>00395                          
    343 <a name="l00404"></a><a class="code" href="classbdm_1_1egamma.html#8a6fd1a1c0190f3e0d95a2a1f99aafc1">00404</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1egamma.html#8a6fd1a1c0190f3e0d95a2a1f99aafc1">from_setting</a>(<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>){ 
    344 <a name="l00405"></a>00405                                 <a class="code" href="classbdm_1_1egamma.html#8a6fd1a1c0190f3e0d95a2a1f99aafc1">epdf::from_setting</a>(<span class="keyword">set</span>); <span class="comment">// reads rv</span> 
    345 <a name="l00406"></a>00406                                 <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>(<a class="code" href="classbdm_1_1egamma.html#0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>,<span class="keyword">set</span>,<span class="stringliteral">"alpha"</span>, UI::compulsory); 
    346 <a name="l00407"></a>00407                                 <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>(beta,<span class="keyword">set</span>,<span class="stringliteral">"beta"</span>, UI::compulsory); 
    347 <a name="l00408"></a>00408                                 <a class="code" href="classbdm_1_1egamma.html#c3a8d8ae8ba79d0fb35036711fa7d127" title="This method TODO.">validate</a>(); 
    348 <a name="l00409"></a>00409                         } 
    349 <a name="l00410"></a><a class="code" href="classbdm_1_1egamma.html#c3a8d8ae8ba79d0fb35036711fa7d127">00410</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1egamma.html#c3a8d8ae8ba79d0fb35036711fa7d127" title="This method TODO.">validate</a>(){ 
    350 <a name="l00411"></a>00411                                 it_assert(<a class="code" href="classbdm_1_1egamma.html#0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>.length() ==beta.length(), <span class="stringliteral">"parameters do not match"</span>); 
    351 <a name="l00412"></a>00412                                 <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> =<a class="code" href="classbdm_1_1egamma.html#0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>.length(); 
    352 <a name="l00413"></a>00413                         } 
    353 <a name="l00414"></a>00414         }; 
    354 <a name="l00415"></a>00415 <a class="code" href="user__info_8h.html#4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a>(egamma); 
    355 <a name="l00432"></a><a class="code" href="classbdm_1_1eigamma.html">00432</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1eigamma.html" title="Inverse-Gamma posterior density.">eigamma</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1egamma.html" title="Gamma posterior density.">egamma</a> 
    356 <a name="l00433"></a>00433         { 
    357 <a name="l00434"></a>00434                 <span class="keyword">protected</span>: 
    358 <a name="l00435"></a>00435                 <span class="keyword">public</span> : 
    359 <a name="l00440"></a>00440  
    360 <a name="l00441"></a><a class="code" href="classbdm_1_1eigamma.html#3aff7bf25ddac27731c60826fcfd878f">00441</a>                         vec <a class="code" href="classbdm_1_1eigamma.html#3aff7bf25ddac27731c60826fcfd878f" title="Returns a sample,  from density .">sample</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> 1.0/<a class="code" href="classbdm_1_1eigamma.html#3aff7bf25ddac27731c60826fcfd878f" title="Returns a sample,  from density .">egamma::sample</a>();}; 
    361 <a name="l00443"></a><a class="code" href="classbdm_1_1eigamma.html#46cecb295edbabd28120cb0f6f572bcb">00443</a>                         vec <a class="code" href="classbdm_1_1eigamma.html#46cecb295edbabd28120cb0f6f572bcb" title="Returns poiter to alpha and beta. Potentially dangerous: use with care!">mean</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> elem_div ( <a class="code" href="classbdm_1_1egamma.html#457bfb1ccb2057df85073e519a15ccc1" title="Vector .">beta</a>,<a class="code" href="classbdm_1_1egamma.html#0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>-1 );} 
    362 <a name="l00444"></a><a class="code" href="classbdm_1_1eigamma.html#c2c696f8c668e9f65392c9449f6a5133">00444</a>                         vec <a class="code" href="classbdm_1_1eigamma.html#c2c696f8c668e9f65392c9449f6a5133" title="return expected variance (not covariance!)">variance</a>()<span class="keyword"> const </span>{vec mea=<a class="code" href="classbdm_1_1eigamma.html#46cecb295edbabd28120cb0f6f572bcb" title="Returns poiter to alpha and beta. Potentially dangerous: use with care!">mean</a>(); <span class="keywordflow">return</span> elem_div ( elem_mult ( mea,mea ),<a class="code" href="classbdm_1_1egamma.html#0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>-2 );} 
    363 <a name="l00445"></a>00445         }; 
    364 <a name="l00446"></a>00446         <span class="comment">/*</span> 
    365 <a name="l00448"></a>00448 <span class="comment">        class emix : public epdf {</span> 
    366 <a name="l00449"></a>00449 <span class="comment">        protected:</span> 
    367 <a name="l00450"></a>00450 <span class="comment">                int n;</span> 
    368 <a name="l00451"></a>00451 <span class="comment">                vec &amp;w;</span> 
    369 <a name="l00452"></a>00452 <span class="comment">                Array&lt;epdf*&gt; Coms;</span> 
    370 <a name="l00453"></a>00453 <span class="comment">        public:</span> 
    371 <a name="l00455"></a>00455 <span class="comment">                emix ( const RV &amp;rv, vec &amp;w0): epdf(rv), n(w0.length()), w(w0), Coms(n) {};</span> 
    372 <a name="l00456"></a>00456 <span class="comment">                void set_parameters( int &amp;i, double wi, epdf* ep){w(i)=wi;Coms(i)=ep;}</span> 
    373 <a name="l00457"></a>00457 <span class="comment">                vec mean(){vec pom; for(int i=0;i&lt;n;i++){pom+=Coms(i)-&gt;mean()*w(i);} return pom;};</span> 
    374 <a name="l00458"></a>00458 <span class="comment">                vec sample() {it_error ( "Not implemented" );return 0;}</span> 
    375 <a name="l00459"></a>00459 <span class="comment">        };</span> 
    376 <a name="l00460"></a>00460 <span class="comment">        */</span> 
    377 <a name="l00461"></a>00461  
    378 <a name="l00463"></a>00463  
    379 <a name="l00464"></a><a class="code" href="classbdm_1_1euni.html">00464</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1euni.html" title="Uniform distributed density on a rectangular support.">euni</a>: <span class="keyword">public</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> 
    380 <a name="l00465"></a>00465         { 
    381 <a name="l00466"></a>00466                 <span class="keyword">protected</span>: 
    382 <a name="l00468"></a><a class="code" href="classbdm_1_1euni.html#ff7ce6a2ef5ef0015bbd1398bed12f32">00468</a>                         vec <a class="code" href="classbdm_1_1euni.html#ff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a>; 
    383 <a name="l00470"></a><a class="code" href="classbdm_1_1euni.html#cfad2dea4a62db6872bda8abd75f0de1">00470</a>                         vec <a class="code" href="classbdm_1_1euni.html#cfad2dea4a62db6872bda8abd75f0de1" title="upper bound on support">high</a>; 
    384 <a name="l00472"></a><a class="code" href="classbdm_1_1euni.html#d3c27e331f90c754d80228108de8ed4c">00472</a>                         vec <a class="code" href="classbdm_1_1euni.html#d3c27e331f90c754d80228108de8ed4c" title="internal">distance</a>; 
    385 <a name="l00474"></a><a class="code" href="classbdm_1_1euni.html#31bb13e8449a8eff35246d46dae35c20">00474</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1euni.html#31bb13e8449a8eff35246d46dae35c20" title="normalizing coefficients">nk</a>; 
    386 <a name="l00476"></a><a class="code" href="classbdm_1_1euni.html#3e63be48dd58659663ca60cd18700476">00476</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1euni.html#3e63be48dd58659663ca60cd18700476" title="cache of log( nk )">lnk</a>; 
    387 <a name="l00477"></a>00477                 <span class="keyword">public</span>: 
    388 <a name="l00480"></a>00480                         <a class="code" href="classbdm_1_1euni.html" title="Uniform distributed density on a rectangular support.">euni</a> ( ) :<a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> ( ) {} 
    389 <a name="l00481"></a>00481                         <a class="code" href="classbdm_1_1euni.html" title="Uniform distributed density on a rectangular support.">euni</a> ( <span class="keyword">const</span> vec &amp;low0, <span class="keyword">const</span> vec &amp;high0 ) {set_parameters ( low0,high0 );} 
    390 <a name="l00482"></a>00482                         <span class="keywordtype">void</span> set_parameters ( <span class="keyword">const</span> vec &amp;low0, <span class="keyword">const</span> vec &amp;high0 ) 
    391 <a name="l00483"></a>00483                         { 
    392 <a name="l00484"></a>00484                                 <a class="code" href="classbdm_1_1euni.html#d3c27e331f90c754d80228108de8ed4c" title="internal">distance</a> = high0-low0; 
    393 <a name="l00485"></a>00485                                 it_assert_debug ( min ( <a class="code" href="classbdm_1_1euni.html#d3c27e331f90c754d80228108de8ed4c" title="internal">distance</a> ) &gt;0.0,<span class="stringliteral">"bad support"</span> ); 
    394 <a name="l00486"></a>00486                                 <a class="code" href="classbdm_1_1euni.html#ff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a> = low0; 
    395 <a name="l00487"></a>00487                                 <a class="code" href="classbdm_1_1euni.html#cfad2dea4a62db6872bda8abd75f0de1" title="upper bound on support">high</a> = high0; 
    396 <a name="l00488"></a>00488                                 <a class="code" href="classbdm_1_1euni.html#31bb13e8449a8eff35246d46dae35c20" title="normalizing coefficients">nk</a> = prod ( 1.0/<a class="code" href="classbdm_1_1euni.html#d3c27e331f90c754d80228108de8ed4c" title="internal">distance</a> ); 
    397 <a name="l00489"></a>00489                                 <a class="code" href="classbdm_1_1euni.html#3e63be48dd58659663ca60cd18700476" title="cache of log( nk )">lnk</a> = log ( <a class="code" href="classbdm_1_1euni.html#31bb13e8449a8eff35246d46dae35c20" title="normalizing coefficients">nk</a> ); 
    398 <a name="l00490"></a>00490                                 <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> = <a class="code" href="classbdm_1_1euni.html#ff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a>.length(); 
    399 <a name="l00491"></a>00491                         } 
    400 <a name="l00493"></a>00493  
    401 <a name="l00494"></a>00494                         <span class="keywordtype">double</span> eval ( <span class="keyword">const</span> vec &amp;val )<span class="keyword"> const  </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1euni.html#31bb13e8449a8eff35246d46dae35c20" title="normalizing coefficients">nk</a>;} 
    402 <a name="l00495"></a><a class="code" href="classbdm_1_1euni.html#caa07b8307bd793d5339d6583e0aba81">00495</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1euni.html#caa07b8307bd793d5339d6583e0aba81">evallog</a> ( <span class="keyword">const</span> vec &amp;val )<span class="keyword"> const  </span>{ 
    403 <a name="l00496"></a>00496                                 <span class="keywordflow">if</span> (any(val&lt;<a class="code" href="classbdm_1_1euni.html#ff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a>) &amp;&amp; any(val&gt;<a class="code" href="classbdm_1_1euni.html#cfad2dea4a62db6872bda8abd75f0de1" title="upper bound on support">high</a>)) {<span class="keywordflow">return</span> inf;} 
    404 <a name="l00497"></a>00497                                 <span class="keywordflow">else</span> <span class="keywordflow">return</span> <a class="code" href="classbdm_1_1euni.html#3e63be48dd58659663ca60cd18700476" title="cache of log( nk )">lnk</a>; 
    405 <a name="l00498"></a>00498                         } 
    406 <a name="l00499"></a><a class="code" href="classbdm_1_1euni.html#fc5df80359ead2918384b2004ce67194">00499</a>                         vec <a class="code" href="classbdm_1_1euni.html#fc5df80359ead2918384b2004ce67194" title="Returns a sample,  from density .">sample</a>()<span class="keyword"> const</span> 
    407 <a name="l00500"></a>00500 <span class="keyword">                        </span>{ 
    408 <a name="l00501"></a>00501                                 vec smp ( <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> ); 
    409 <a name="l00502"></a>00502 <span class="preprocessor">#pragma omp critical</span> 
    410 <a name="l00503"></a>00503 <span class="preprocessor"></span>                                UniRNG.sample_vector ( <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> ,smp ); 
    411 <a name="l00504"></a>00504                                 <span class="keywordflow">return</span> <a class="code" href="classbdm_1_1euni.html#ff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a>+elem_mult ( <a class="code" href="classbdm_1_1euni.html#d3c27e331f90c754d80228108de8ed4c" title="internal">distance</a>,smp ); 
    412 <a name="l00505"></a>00505                         } 
    413 <a name="l00507"></a><a class="code" href="classbdm_1_1euni.html#46caa8c13aba2e6228f964208918b226">00507</a>                         vec <a class="code" href="classbdm_1_1euni.html#46caa8c13aba2e6228f964208918b226" title="set values of low and high ">mean</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> ( <a class="code" href="classbdm_1_1euni.html#cfad2dea4a62db6872bda8abd75f0de1" title="upper bound on support">high</a>-<a class="code" href="classbdm_1_1euni.html#ff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a> ) /2.0;} 
    414 <a name="l00508"></a><a class="code" href="classbdm_1_1euni.html#951f932155111f6053c980f672b4c22c">00508</a>                         vec <a class="code" href="classbdm_1_1euni.html#951f932155111f6053c980f672b4c22c" title="return expected variance (not covariance!)">variance</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> ( pow ( <a class="code" href="classbdm_1_1euni.html#cfad2dea4a62db6872bda8abd75f0de1" title="upper bound on support">high</a>,2 ) +pow ( <a class="code" href="classbdm_1_1euni.html#ff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a>,2 ) +elem_mult ( <a class="code" href="classbdm_1_1euni.html#cfad2dea4a62db6872bda8abd75f0de1" title="upper bound on support">high</a>,<a class="code" href="classbdm_1_1euni.html#ff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a> ) ) /3.0;} 
    415 <a name="l00517"></a><a class="code" href="classbdm_1_1euni.html#77f5fef1f006fe056066da23b9e5f042">00517</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1euni.html#77f5fef1f006fe056066da23b9e5f042">from_setting</a>(<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>){ 
    416 <a name="l00518"></a>00518                                 <a class="code" href="classbdm_1_1euni.html#77f5fef1f006fe056066da23b9e5f042">epdf::from_setting</a>(<span class="keyword">set</span>); <span class="comment">// reads rv and rvc</span> 
    417 <a name="l00519"></a>00519  
    418 <a name="l00520"></a>00520                                 <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>(<a class="code" href="classbdm_1_1euni.html#cfad2dea4a62db6872bda8abd75f0de1" title="upper bound on support">high</a>,<span class="keyword">set</span>,<span class="stringliteral">"high"</span>, UI::compulsory); 
    419 <a name="l00521"></a>00521                                 <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>(<a class="code" href="classbdm_1_1euni.html#ff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a>,<span class="keyword">set</span>,<span class="stringliteral">"low"</span>, UI::compulsory); 
    420 <a name="l00522"></a>00522                         } 
    421 <a name="l00523"></a>00523         }; 
     232<a name="l00233"></a>00233                 ldmat <a class="code" href="classbdm_1_1egiw.html#a88c321a2051d1afdbb31a098896a717b" title="Covariance of the LS estimate.">est_theta_cov</a>() <span class="keyword">const</span>; 
     233<a name="l00234"></a>00234  
     234<a name="l00236"></a>00236                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1egiw.html#ad2075aa2306648b3e4fe40bb86628d5c" title="expected values of the linear coefficient and the covariance matrix are written to...">mean_mat</a> (mat &amp;M, mat&amp;R) <span class="keyword">const</span>; 
     235<a name="l00238"></a>00238                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1egiw.html#abfb8e7c619b34ad804a73bff71742b5e" title="In this instance, val= [theta, r]. For multivariate instances, it is stored columnwise...">evallog_nn</a> (<span class="keyword">const</span> vec &amp;val) <span class="keyword">const</span>; 
     236<a name="l00239"></a>00239                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1egiw.html#a41d72ba7b2abc8a9a4209ffa98ed5633" title="logarithm of the normalizing constant, ">lognc</a> () <span class="keyword">const</span>; 
     237<a name="l00240"></a><a class="code" href="classbdm_1_1egiw.html#a8e610e95401a11baf34f65e16ecd87be">00240</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1egiw.html#a8e610e95401a11baf34f65e16ecd87be" title="Power of the density, used e.g. to flatten the density.">pow</a> (<span class="keywordtype">double</span> p) {V *= p;<a class="code" href="classbdm_1_1egiw.html#a447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a> *= p;}; 
     238<a name="l00241"></a>00241  
     239<a name="l00244"></a>00244  
     240<a name="l00245"></a>00245                 <a class="code" href="classbdm_1_1ldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a>&amp; _V() {<span class="keywordflow">return</span> V;} 
     241<a name="l00246"></a>00246                 <span class="keyword">const</span> ldmat&amp; _V()<span class="keyword"> const </span>{<span class="keywordflow">return</span> V;} 
     242<a name="l00247"></a>00247                 <span class="keywordtype">double</span>&amp; _nu()  {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1egiw.html#a447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a>;} 
     243<a name="l00248"></a>00248                 <span class="keyword">const</span> <span class="keywordtype">double</span>&amp; _nu()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1egiw.html#a447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a>;} 
     244<a name="l00249"></a><a class="code" href="classbdm_1_1egiw.html#a55b76ec75bd2df5ef9cab3be20a33bbb">00249</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1egiw.html#a55b76ec75bd2df5ef9cab3be20a33bbb">from_setting</a> (<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>) { 
     245<a name="l00250"></a>00250                         <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (<a class="code" href="classbdm_1_1egiw.html#a447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a>, <span class="keyword">set</span>, <span class="stringliteral">&quot;nu&quot;</span>, UI::compulsory); 
     246<a name="l00251"></a>00251                         <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (<a class="code" href="classbdm_1_1egiw.html#a23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a>, <span class="keyword">set</span>, <span class="stringliteral">&quot;dimx&quot;</span>, UI::compulsory); 
     247<a name="l00252"></a>00252                         mat V; 
     248<a name="l00253"></a>00253                         <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (V, <span class="keyword">set</span>, <span class="stringliteral">&quot;V&quot;</span>, UI::compulsory); 
     249<a name="l00254"></a>00254                         set_parameters (<a class="code" href="classbdm_1_1egiw.html#a23e4d78bea7e98840f3da30e76a2b57a" title="Dimension of the output.">dimx</a>, V, <a class="code" href="classbdm_1_1egiw.html#a447eacf19d4f4083872686f044814dc4" title="Number of data records (degrees of freedom) of sufficient statistics.">nu</a>); 
     250<a name="l00255"></a>00255                         <a class="code" href="classbdm_1_1shared__ptr.html" title="A naive implementation of roughly a subset of the std::tr1::shared_ptr spec.">shared_ptr&lt;RV&gt;</a> <a class="code" href="classbdm_1_1epdf.html#a62c5b8ff71d9ebe6cd58d3c342eb1dc8" title="Description of the random variable.">rv</a> = UI::build&lt;RV&gt; (<span class="keyword">set</span>, <span class="stringliteral">&quot;rv&quot;</span>, UI::compulsory); 
     251<a name="l00256"></a>00256                         <a class="code" href="classbdm_1_1epdf.html#af423e28448dbb69ef4905295ec8de8ff" title="Name its rv.">set_rv</a> (*rv); 
     252<a name="l00257"></a>00257                 } 
     253<a name="l00259"></a>00259 }; 
     254<a name="l00260"></a>00260 <a class="code" href="user__info_8h.html#a4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a> ( egiw ); 
     255<a name="l00261"></a>00261 SHAREDPTR ( egiw ); 
     256<a name="l00262"></a>00262  
     257<a name="l00271"></a><a class="code" href="classbdm_1_1eDirich.html">00271</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1eDirich.html" title="Dirichlet posterior density.">eDirich</a>: <span class="keyword">public</span> <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> 
     258<a name="l00272"></a>00272 { 
     259<a name="l00273"></a>00273         <span class="keyword">protected</span>: 
     260<a name="l00275"></a><a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2">00275</a>                 vec <a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>; 
     261<a name="l00276"></a>00276         <span class="keyword">public</span>: 
     262<a name="l00279"></a>00279  
     263<a name="l00280"></a>00280                 <a class="code" href="classbdm_1_1eDirich.html" title="Dirichlet posterior density.">eDirich</a> () : <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> () {}; 
     264<a name="l00281"></a>00281                 <a class="code" href="classbdm_1_1eDirich.html" title="Dirichlet posterior density.">eDirich</a> (<span class="keyword">const</span> <a class="code" href="classbdm_1_1eDirich.html" title="Dirichlet posterior density.">eDirich</a> &amp;D0) : <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> () {set_parameters (D0.<a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>);}; 
     265<a name="l00282"></a>00282                 eDirich (<span class="keyword">const</span> vec &amp;beta0) {set_parameters (beta0);}; 
     266<a name="l00283"></a>00283                 <span class="keywordtype">void</span> set_parameters (<span class="keyword">const</span> vec &amp;beta0) { 
     267<a name="l00284"></a>00284                         <a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a> = beta0; 
     268<a name="l00285"></a>00285                         <a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> = <a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>.length(); 
     269<a name="l00286"></a>00286                 } 
     270<a name="l00288"></a>00288  
     271<a name="l00289"></a><a class="code" href="classbdm_1_1eDirich.html#a3290613d31d58daa8a45a54b003871fc">00289</a>                 vec <a class="code" href="classbdm_1_1eDirich.html#a3290613d31d58daa8a45a54b003871fc" title="Returns a sample,  from density .">sample</a>()<span class="keyword"> const </span>{ 
     272<a name="l00290"></a>00290                         <a class="code" href="bdmerror_8h.html#a7c43f3a72afe68ab0c85663a1bb3521a" title="Unconditionally throw std::runtime_error.">bdm_error</a> (<span class="stringliteral">&quot;Not implemented&quot;</span>); 
     273<a name="l00291"></a>00291                         <span class="keywordflow">return</span> vec(); 
     274<a name="l00292"></a>00292                 } 
     275<a name="l00293"></a>00293  
     276<a name="l00294"></a><a class="code" href="classbdm_1_1eDirich.html#acb343355ec791298bb5a3404cd482fb6">00294</a>                 vec <a class="code" href="classbdm_1_1eDirich.html#acb343355ec791298bb5a3404cd482fb6" title="return expected value">mean</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a> / sum (<a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>);}; 
     277<a name="l00295"></a><a class="code" href="classbdm_1_1eDirich.html#a43c547a2507e233706f92712d8c2aacc">00295</a>                 vec <a class="code" href="classbdm_1_1eDirich.html#a43c547a2507e233706f92712d8c2aacc" title="return expected variance (not covariance!)">variance</a>()<span class="keyword"> const </span>{<span class="keywordtype">double</span> gamma = sum (<a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>); <span class="keywordflow">return</span> elem_mult (<a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>, (<a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a> + 1)) / (gamma* (gamma + 1));} 
     278<a name="l00297"></a><a class="code" href="classbdm_1_1eDirich.html#ae09a24938e80c3d94b0ee842d1552318">00297</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eDirich.html#ae09a24938e80c3d94b0ee842d1552318" title="In this instance, val is ...">evallog_nn</a> (<span class="keyword">const</span> vec &amp;val)<span class="keyword"> const </span>{ 
     279<a name="l00298"></a>00298                         <span class="keywordtype">double</span> tmp; tmp = (<a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a> - 1) * log (val); 
     280<a name="l00299"></a>00299                         <span class="keywordflow">return</span> tmp; 
     281<a name="l00300"></a>00300                 } 
     282<a name="l00301"></a>00301  
     283<a name="l00302"></a><a class="code" href="classbdm_1_1eDirich.html#a279a99f6266c82fe2273e83841f19eb2">00302</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eDirich.html#a279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a> ()<span class="keyword"> const </span>{ 
     284<a name="l00303"></a>00303                         <span class="keywordtype">double</span> tmp; 
     285<a name="l00304"></a>00304                         <span class="keywordtype">double</span> gam = sum (<a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>); 
     286<a name="l00305"></a>00305                         <span class="keywordtype">double</span> lgb = 0.0; 
     287<a name="l00306"></a>00306                         <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = 0;i &lt; <a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>.length();i++) {lgb += lgamma (<a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a> (i));} 
     288<a name="l00307"></a>00307                         tmp = lgb - lgamma (gam); 
     289<a name="l00308"></a>00308                         <span class="keywordflow">return</span> tmp; 
     290<a name="l00309"></a>00309                 } 
     291<a name="l00310"></a>00310  
     292<a name="l00312"></a><a class="code" href="classbdm_1_1eDirich.html#a175e0add26d2105c28d8121eefb9e324">00312</a>                 vec&amp; <a class="code" href="classbdm_1_1eDirich.html#a175e0add26d2105c28d8121eefb9e324" title="access function">_beta</a>()  {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1eDirich.html#af25886a49b4667af61245de81c83b5d2" title="sufficient statistics">beta</a>;} 
     293<a name="l00314"></a>00314 }; 
     294<a name="l00315"></a>00315  
     295<a name="l00317"></a><a class="code" href="classbdm_1_1multiBM.html">00317</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1multiBM.html" title="Estimator for Multinomial density.">multiBM</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a> 
     296<a name="l00318"></a>00318 { 
     297<a name="l00319"></a>00319         <span class="keyword">protected</span>: 
     298<a name="l00321"></a><a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a">00321</a>                 <a class="code" href="classbdm_1_1eDirich.html" title="Dirichlet posterior density.">eDirich</a> <a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>; 
     299<a name="l00323"></a><a class="code" href="classbdm_1_1multiBM.html#a044263356944c92209eecd39a5187d25">00323</a>                 vec &amp;<a class="code" href="classbdm_1_1multiBM.html#a044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a>; 
     300<a name="l00324"></a>00324         <span class="keyword">public</span>: 
     301<a name="l00326"></a><a class="code" href="classbdm_1_1multiBM.html#ac4dd6d9522a8a605776d21bac9bd9daf">00326</a>                 <a class="code" href="classbdm_1_1multiBM.html#ac4dd6d9522a8a605776d21bac9bd9daf" title="Default constructor.">multiBM</a> () : <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a> (), <a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a> (), <a class="code" href="classbdm_1_1multiBM.html#a044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a> (<a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>._beta()) { 
     302<a name="l00327"></a>00327                         <span class="keywordflow">if</span> (<a class="code" href="classbdm_1_1multiBM.html#a044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a>.length() &gt; 0) {<a class="code" href="classbdm_1_1BMEF.html#a06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a> = <a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>.<a class="code" href="classbdm_1_1eDirich.html#a279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>();} 
     303<a name="l00328"></a>00328                         <span class="keywordflow">else</span>{<a class="code" href="classbdm_1_1BMEF.html#a06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a> = 0.0;} 
     304<a name="l00329"></a>00329                 } 
     305<a name="l00331"></a><a class="code" href="classbdm_1_1multiBM.html#ac4378cf8037f6bed29c74eea63344b31">00331</a>                 <a class="code" href="classbdm_1_1multiBM.html#ac4378cf8037f6bed29c74eea63344b31" title="Copy constructor.">multiBM</a> (<span class="keyword">const</span> <a class="code" href="classbdm_1_1multiBM.html" title="Estimator for Multinomial density.">multiBM</a> &amp;B) : <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a> (B), <a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a> (B.<a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>), <a class="code" href="classbdm_1_1multiBM.html#a044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a> (<a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>._beta()) {} 
     306<a name="l00333"></a><a class="code" href="classbdm_1_1multiBM.html#adbe6b90d410dc062a233d1dc09eeba52">00333</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1multiBM.html#adbe6b90d410dc062a233d1dc09eeba52" title="Sets sufficient statistics to match that of givefrom mB0.">set_statistics</a> (<span class="keyword">const</span> <a class="code" href="classbdm_1_1BM.html" title="Bayesian Model of a system, i.e. all uncertainty is modeled by probabilities.">BM</a>* mB0) {<span class="keyword">const</span> <a class="code" href="classbdm_1_1multiBM.html" title="Estimator for Multinomial density.">multiBM</a>* mB = <span class="keyword">dynamic_cast&lt;</span><span class="keyword">const </span><a class="code" href="classbdm_1_1multiBM.html" title="Estimator for Multinomial density.">multiBM</a>*<span class="keyword">&gt;</span> (mB0); <a class="code" href="classbdm_1_1multiBM.html#a044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a> = mB-&gt;<a class="code" href="classbdm_1_1multiBM.html#a044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a>;} 
     307<a name="l00334"></a><a class="code" href="classbdm_1_1multiBM.html#a1e4bf41b61937fd80f34049742e23f95">00334</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1multiBM.html#a1e4bf41b61937fd80f34049742e23f95" title="Incremental Bayes rule.">bayes</a> (<span class="keyword">const</span> vec &amp;dt) { 
     308<a name="l00335"></a>00335                         <span class="keywordflow">if</span> (<a class="code" href="classbdm_1_1BMEF.html#a1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a> &lt; 1.0) {<a class="code" href="classbdm_1_1multiBM.html#a044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a> *= <a class="code" href="classbdm_1_1BMEF.html#a1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a>;<a class="code" href="classbdm_1_1BMEF.html#a06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a> = <a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>.<a class="code" href="classbdm_1_1eDirich.html#a279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>();} 
     309<a name="l00336"></a>00336                         <a class="code" href="classbdm_1_1multiBM.html#a044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a> += dt; 
     310<a name="l00337"></a>00337                         <span class="keywordflow">if</span> (<a class="code" href="classbdm_1_1BM.html#afaff0ad12556fe7dc0e2807d4fd938ee" title="If true, the filter will compute likelihood of the data record and store it in ll...">evalll</a>) {<a class="code" href="classbdm_1_1BM.html#a4064b6559d962633e4372b12f4cd204a" title="Logarithm of marginalized data likelihood.">ll</a> = <a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>.<a class="code" href="classbdm_1_1eDirich.html#a279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>() - <a class="code" href="classbdm_1_1BMEF.html#a06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a>;} 
     311<a name="l00338"></a>00338                 } 
     312<a name="l00339"></a><a class="code" href="classbdm_1_1multiBM.html#ae157b607c1e3fa91d42aeea44458e2bf">00339</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1multiBM.html#ae157b607c1e3fa91d42aeea44458e2bf">logpred</a> (<span class="keyword">const</span> vec &amp;dt)<span class="keyword"> const </span>{ 
     313<a name="l00340"></a>00340                         <a class="code" href="classbdm_1_1eDirich.html" title="Dirichlet posterior density.">eDirich</a> pred (<a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>); 
     314<a name="l00341"></a>00341                         vec &amp;<a class="code" href="classbdm_1_1multiBM.html#a044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a> = pred.<a class="code" href="classbdm_1_1eDirich.html#a175e0add26d2105c28d8121eefb9e324" title="access function">_beta</a>(); 
     315<a name="l00342"></a>00342  
     316<a name="l00343"></a>00343                         <span class="keywordtype">double</span> lll; 
     317<a name="l00344"></a>00344                         <span class="keywordflow">if</span> (<a class="code" href="classbdm_1_1BMEF.html#a1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a> &lt; 1.0) 
     318<a name="l00345"></a>00345                                 {beta *= <a class="code" href="classbdm_1_1BMEF.html#a1331865e10fb1ccef65bb4c47fa3be64" title="forgetting factor">frg</a>;lll = pred.<a class="code" href="classbdm_1_1eDirich.html#a279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>();} 
     319<a name="l00346"></a>00346                         <span class="keywordflow">else</span> 
     320<a name="l00347"></a>00347                                 <span class="keywordflow">if</span> (<a class="code" href="classbdm_1_1BM.html#afaff0ad12556fe7dc0e2807d4fd938ee" title="If true, the filter will compute likelihood of the data record and store it in ll...">evalll</a>) {lll = <a class="code" href="classbdm_1_1BMEF.html#a06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a>;} 
     321<a name="l00348"></a>00348                                 <span class="keywordflow">else</span>{lll = pred.<a class="code" href="classbdm_1_1eDirich.html#a279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>();} 
     322<a name="l00349"></a>00349  
     323<a name="l00350"></a>00350                         beta += dt; 
     324<a name="l00351"></a>00351                         <span class="keywordflow">return</span> pred.<a class="code" href="classbdm_1_1eDirich.html#a279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>() - lll; 
     325<a name="l00352"></a>00352                 } 
     326<a name="l00353"></a>00353                 <span class="keywordtype">void</span> flatten (<span class="keyword">const</span> <a class="code" href="classbdm_1_1BMEF.html" title="Estimator for Exponential family.">BMEF</a>* B) { 
     327<a name="l00354"></a>00354                         <span class="keyword">const</span> <a class="code" href="classbdm_1_1multiBM.html" title="Estimator for Multinomial density.">multiBM</a>* E = <span class="keyword">dynamic_cast&lt;</span><span class="keyword">const </span><a class="code" href="classbdm_1_1multiBM.html" title="Estimator for Multinomial density.">multiBM</a>*<span class="keyword">&gt;</span> (B); 
     328<a name="l00355"></a>00355                         <span class="comment">// sum(beta) should be equal to sum(B.beta)</span> 
     329<a name="l00356"></a>00356                         <span class="keyword">const</span> vec &amp;Eb = E-&gt;<a class="code" href="classbdm_1_1multiBM.html#a044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a>;<span class="comment">//const_cast&lt;multiBM*&gt; ( E )-&gt;_beta();</span> 
     330<a name="l00357"></a>00357                         <a class="code" href="classbdm_1_1multiBM.html#a044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a> *= (sum (Eb) / sum (<a class="code" href="classbdm_1_1multiBM.html#a044263356944c92209eecd39a5187d25" title="Pointer inside est to sufficient statistics.">beta</a>)); 
     331<a name="l00358"></a>00358                         <span class="keywordflow">if</span> (<a class="code" href="classbdm_1_1BM.html#afaff0ad12556fe7dc0e2807d4fd938ee" title="If true, the filter will compute likelihood of the data record and store it in ll...">evalll</a>) {<a class="code" href="classbdm_1_1BMEF.html#a06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a> = <a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>.<a class="code" href="classbdm_1_1eDirich.html#a279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>();} 
     332<a name="l00359"></a>00359                 } 
     333<a name="l00361"></a><a class="code" href="classbdm_1_1multiBM.html#a31ff93f89473f099e489b9e1dc8d9513">00361</a>                 <span class="keyword">const</span> <a class="code" href="classbdm_1_1eDirich.html" title="Dirichlet posterior density.">eDirich</a>&amp; <a class="code" href="classbdm_1_1multiBM.html#a31ff93f89473f099e489b9e1dc8d9513" title="reimplemnetation of BM::posterior()">posterior</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>;}; 
     334<a name="l00363"></a><a class="code" href="classbdm_1_1multiBM.html#a7a480eace4446661bacca94c57499f01">00363</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1multiBM.html#a7a480eace4446661bacca94c57499f01" title="constructor function">set_parameters</a> (<span class="keyword">const</span> vec &amp;beta0) { 
     335<a name="l00364"></a>00364                         <a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>.set_parameters (beta0); 
     336<a name="l00365"></a>00365                         <span class="keywordflow">if</span> (<a class="code" href="classbdm_1_1BM.html#afaff0ad12556fe7dc0e2807d4fd938ee" title="If true, the filter will compute likelihood of the data record and store it in ll...">evalll</a>) {<a class="code" href="classbdm_1_1BMEF.html#a06e7b3ac03e10017d4288c76888e2865" title="cached value of lognc() in the previous step (used in evaluation of ll )">last_lognc</a> = <a class="code" href="classbdm_1_1multiBM.html#a9ecc6878abbd20eb8d8e43b6ab3f941a" title="Conjugate prior and posterior.">est</a>.<a class="code" href="classbdm_1_1eDirich.html#a279a99f6266c82fe2273e83841f19eb2" title="logarithm of the normalizing constant, ">lognc</a>();} 
     337<a name="l00366"></a>00366                 } 
     338<a name="l00367"></a>00367 }; 
     339<a name="l00368"></a>00368  
     340<a name="l00378"></a><a class="code" href="classbdm_1_1egamma.html">00378</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1egamma.html" title="Gamma posterior density.">egamma</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> 
     341<a name="l00379"></a>00379 { 
     342<a name="l00380"></a>00380         <span class="keyword">protected</span>: 
     343<a name="l00382"></a><a class="code" href="classbdm_1_1egamma.html#a0901ec983e66b8337aaa506e13b122fa">00382</a>                 vec <a class="code" href="classbdm_1_1egamma.html#a0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>; 
     344<a name="l00384"></a><a class="code" href="classbdm_1_1egamma.html#a457bfb1ccb2057df85073e519a15ccc1">00384</a>                 vec <a class="code" href="classbdm_1_1egamma.html#a457bfb1ccb2057df85073e519a15ccc1" title="Vector .">beta</a>; 
     345<a name="l00385"></a>00385         <span class="keyword">public</span> : 
     346<a name="l00388"></a>00388                 <a class="code" href="classbdm_1_1egamma.html" title="Gamma posterior density.">egamma</a> () : <a class="code" href="classbdm_1_1eEF.html" title="General conjugate exponential family posterior density.">eEF</a> (), <a class="code" href="classbdm_1_1egamma.html#a0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a> (0), <a class="code" href="classbdm_1_1egamma.html#a457bfb1ccb2057df85073e519a15ccc1" title="Vector .">beta</a> (0) {}; 
     347<a name="l00389"></a>00389                 <a class="code" href="classbdm_1_1egamma.html" title="Gamma posterior density.">egamma</a> (<span class="keyword">const</span> vec &amp;a, <span class="keyword">const</span> vec &amp;b) {set_parameters (a, b);}; 
     348<a name="l00390"></a>00390                 <span class="keywordtype">void</span> set_parameters (<span class="keyword">const</span> vec &amp;a, <span class="keyword">const</span> vec &amp;b) {<a class="code" href="classbdm_1_1egamma.html#a0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a> = a, <a class="code" href="classbdm_1_1egamma.html#a457bfb1ccb2057df85073e519a15ccc1" title="Vector .">beta</a> = b;<a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> = <a class="code" href="classbdm_1_1egamma.html#a0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>.length();}; 
     349<a name="l00392"></a>00392  
     350<a name="l00393"></a>00393                 vec <a class="code" href="classbdm_1_1egamma.html#a6ed82f0fd05f6002487256d8e75a0bbd" title="Returns a sample,  from density .">sample</a>() <span class="keyword">const</span>; 
     351<a name="l00394"></a>00394                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1egamma.html#aa8e11e5a580ff42a1b205974c60768c6" title="Evaluate normalized log-probability.">evallog</a> (<span class="keyword">const</span> vec &amp;val) <span class="keyword">const</span>; 
     352<a name="l00395"></a>00395                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1egamma.html#a9a66cbd100e8520c769ccb3c451f86f8" title="logarithm of the normalizing constant, ">lognc</a> () <span class="keyword">const</span>; 
     353<a name="l00397"></a><a class="code" href="classbdm_1_1egamma.html#a0865cb3d6339fdc7410806cf70a329ed">00397</a>                 vec&amp; <a class="code" href="classbdm_1_1egamma.html#a0865cb3d6339fdc7410806cf70a329ed" title="Returns pointer to internal alpha. Potentially dengerous: use with care!">_alpha</a>() {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1egamma.html#a0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>;} 
     354<a name="l00399"></a><a class="code" href="classbdm_1_1egamma.html#ac42cadd9cbd344caaa69b0b433cd16ca">00399</a>                 vec&amp; <a class="code" href="classbdm_1_1egamma.html#ac42cadd9cbd344caaa69b0b433cd16ca" title="Returns pointer to internal beta. Potentially dengerous: use with care!">_beta</a>() {<span class="keywordflow">return</span> beta;} 
     355<a name="l00400"></a><a class="code" href="classbdm_1_1egamma.html#a49d256c42cce14c6faa56ec242b57e85">00400</a>                 vec <a class="code" href="classbdm_1_1egamma.html#a49d256c42cce14c6faa56ec242b57e85" title="return expected value">mean</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> elem_div (<a class="code" href="classbdm_1_1egamma.html#a0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>, beta);} 
     356<a name="l00401"></a><a class="code" href="classbdm_1_1egamma.html#a36986cc01917cd0796fadc17125bdec1">00401</a>                 vec <a class="code" href="classbdm_1_1egamma.html#a36986cc01917cd0796fadc17125bdec1" title="return expected variance (not covariance!)">variance</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> elem_div (<a class="code" href="classbdm_1_1egamma.html#a0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>, elem_mult (beta, beta)); } 
     357<a name="l00402"></a>00402  
     358<a name="l00411"></a><a class="code" href="classbdm_1_1egamma.html#a8a6fd1a1c0190f3e0d95a2a1f99aafc1">00411</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1egamma.html#a8a6fd1a1c0190f3e0d95a2a1f99aafc1">from_setting</a> (<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>) { 
     359<a name="l00412"></a>00412                         <a class="code" href="classbdm_1_1egamma.html#a8a6fd1a1c0190f3e0d95a2a1f99aafc1">epdf::from_setting</a> (<span class="keyword">set</span>); <span class="comment">// reads rv</span> 
     360<a name="l00413"></a>00413                         <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (<a class="code" href="classbdm_1_1egamma.html#a0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>, <span class="keyword">set</span>, <span class="stringliteral">&quot;alpha&quot;</span>, UI::compulsory); 
     361<a name="l00414"></a>00414                         <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (beta, <span class="keyword">set</span>, <span class="stringliteral">&quot;beta&quot;</span>, UI::compulsory); 
     362<a name="l00415"></a>00415                         <a class="code" href="classbdm_1_1egamma.html#ac3a8d8ae8ba79d0fb35036711fa7d127" title="This method TODO.">validate</a>(); 
     363<a name="l00416"></a>00416                 } 
     364<a name="l00417"></a><a class="code" href="classbdm_1_1egamma.html#ac3a8d8ae8ba79d0fb35036711fa7d127">00417</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1egamma.html#ac3a8d8ae8ba79d0fb35036711fa7d127" title="This method TODO.">validate</a>() { 
     365<a name="l00418"></a>00418                         <a class="code" href="bdmerror_8h.html#a89a0f906b242b79c5d3d342291b2cab4" title="Throw std::runtime_exception if t is not true and NDEBUG is not defined.">bdm_assert_debug</a> (<a class="code" href="classbdm_1_1egamma.html#a0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>.length() == beta.length(), <span class="stringliteral">&quot;parameters do not match&quot;</span>); 
     366<a name="l00419"></a>00419                         <a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> = <a class="code" href="classbdm_1_1egamma.html#a0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a>.length(); 
     367<a name="l00420"></a>00420                 } 
     368<a name="l00421"></a>00421 }; 
     369<a name="l00422"></a>00422 <a class="code" href="user__info_8h.html#a4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a> (egamma); 
     370<a name="l00423"></a>00423 SHAREDPTR ( egamma ); 
     371<a name="l00424"></a>00424  
     372<a name="l00441"></a><a class="code" href="classbdm_1_1eigamma.html">00441</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1eigamma.html" title="Inverse-Gamma posterior density.">eigamma</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1egamma.html" title="Gamma posterior density.">egamma</a> 
     373<a name="l00442"></a>00442 { 
     374<a name="l00443"></a>00443         <span class="keyword">protected</span>: 
     375<a name="l00444"></a>00444         <span class="keyword">public</span> : 
     376<a name="l00449"></a>00449  
     377<a name="l00450"></a><a class="code" href="classbdm_1_1eigamma.html#a3aff7bf25ddac27731c60826fcfd878f">00450</a>                 vec <a class="code" href="classbdm_1_1eigamma.html#a3aff7bf25ddac27731c60826fcfd878f" title="Returns a sample,  from density .">sample</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> 1.0 / <a class="code" href="classbdm_1_1eigamma.html#a3aff7bf25ddac27731c60826fcfd878f" title="Returns a sample,  from density .">egamma::sample</a>();}; 
     378<a name="l00452"></a><a class="code" href="classbdm_1_1eigamma.html#a46cecb295edbabd28120cb0f6f572bcb">00452</a>                 vec <a class="code" href="classbdm_1_1eigamma.html#a46cecb295edbabd28120cb0f6f572bcb" title="Returns poiter to alpha and beta. Potentially dangerous: use with care!">mean</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> elem_div (<a class="code" href="classbdm_1_1egamma.html#a457bfb1ccb2057df85073e519a15ccc1" title="Vector .">beta</a>, <a class="code" href="classbdm_1_1egamma.html#a0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a> - 1);} 
     379<a name="l00453"></a><a class="code" href="classbdm_1_1eigamma.html#ac2c696f8c668e9f65392c9449f6a5133">00453</a>                 vec <a class="code" href="classbdm_1_1eigamma.html#ac2c696f8c668e9f65392c9449f6a5133" title="return expected variance (not covariance!)">variance</a>()<span class="keyword"> const </span>{vec mea = <a class="code" href="classbdm_1_1eigamma.html#a46cecb295edbabd28120cb0f6f572bcb" title="Returns poiter to alpha and beta. Potentially dangerous: use with care!">mean</a>(); <span class="keywordflow">return</span> elem_div (elem_mult (mea, mea), <a class="code" href="classbdm_1_1egamma.html#a0901ec983e66b8337aaa506e13b122fa" title="Vector .">alpha</a> - 2);} 
     380<a name="l00454"></a>00454 }; 
     381<a name="l00455"></a>00455 <span class="comment">/*</span> 
     382<a name="l00457"></a>00457 <span class="comment">class emix : public epdf {</span> 
     383<a name="l00458"></a>00458 <span class="comment">protected:</span> 
     384<a name="l00459"></a>00459 <span class="comment">        int n;</span> 
     385<a name="l00460"></a>00460 <span class="comment">        vec &amp;w;</span> 
     386<a name="l00461"></a>00461 <span class="comment">        Array&lt;epdf*&gt; Coms;</span> 
     387<a name="l00462"></a>00462 <span class="comment">public:</span> 
     388<a name="l00464"></a>00464 <span class="comment">        emix ( const RV &amp;rv, vec &amp;w0): epdf(rv), n(w0.length()), w(w0), Coms(n) {};</span> 
     389<a name="l00465"></a>00465 <span class="comment">        void set_parameters( int &amp;i, double wi, epdf* ep){w(i)=wi;Coms(i)=ep;}</span> 
     390<a name="l00466"></a>00466 <span class="comment">        vec mean(){vec pom; for(int i=0;i&lt;n;i++){pom+=Coms(i)-&gt;mean()*w(i);} return pom;};</span> 
     391<a name="l00467"></a>00467 <span class="comment">};</span> 
     392<a name="l00468"></a>00468 <span class="comment">*/</span> 
     393<a name="l00469"></a>00469  
     394<a name="l00471"></a>00471  
     395<a name="l00472"></a><a class="code" href="classbdm_1_1euni.html">00472</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1euni.html" title="Uniform distributed density on a rectangular support.">euni</a>: <span class="keyword">public</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> 
     396<a name="l00473"></a>00473 { 
     397<a name="l00474"></a>00474         <span class="keyword">protected</span>: 
     398<a name="l00476"></a><a class="code" href="classbdm_1_1euni.html#aff7ce6a2ef5ef0015bbd1398bed12f32">00476</a>                 vec <a class="code" href="classbdm_1_1euni.html#aff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a>; 
     399<a name="l00478"></a><a class="code" href="classbdm_1_1euni.html#acfad2dea4a62db6872bda8abd75f0de1">00478</a>                 vec <a class="code" href="classbdm_1_1euni.html#acfad2dea4a62db6872bda8abd75f0de1" title="upper bound on support">high</a>; 
     400<a name="l00480"></a><a class="code" href="classbdm_1_1euni.html#ad3c27e331f90c754d80228108de8ed4c">00480</a>                 vec <a class="code" href="classbdm_1_1euni.html#ad3c27e331f90c754d80228108de8ed4c" title="internal">distance</a>; 
     401<a name="l00482"></a><a class="code" href="classbdm_1_1euni.html#a31bb13e8449a8eff35246d46dae35c20">00482</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1euni.html#a31bb13e8449a8eff35246d46dae35c20" title="normalizing coefficients">nk</a>; 
     402<a name="l00484"></a><a class="code" href="classbdm_1_1euni.html#a3e63be48dd58659663ca60cd18700476">00484</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1euni.html#a3e63be48dd58659663ca60cd18700476" title="cache of log( nk )">lnk</a>; 
     403<a name="l00485"></a>00485         <span class="keyword">public</span>: 
     404<a name="l00488"></a>00488                 <a class="code" href="classbdm_1_1euni.html" title="Uniform distributed density on a rectangular support.">euni</a> () : <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> () {} 
     405<a name="l00489"></a>00489                 <a class="code" href="classbdm_1_1euni.html" title="Uniform distributed density on a rectangular support.">euni</a> (<span class="keyword">const</span> vec &amp;low0, <span class="keyword">const</span> vec &amp;high0) {set_parameters (low0, high0);} 
     406<a name="l00490"></a>00490                 <span class="keywordtype">void</span> set_parameters (<span class="keyword">const</span> vec &amp;low0, <span class="keyword">const</span> vec &amp;high0) { 
     407<a name="l00491"></a>00491                         <a class="code" href="classbdm_1_1euni.html#ad3c27e331f90c754d80228108de8ed4c" title="internal">distance</a> = high0 - low0; 
     408<a name="l00492"></a>00492                         <a class="code" href="bdmerror_8h.html#a89a0f906b242b79c5d3d342291b2cab4" title="Throw std::runtime_exception if t is not true and NDEBUG is not defined.">bdm_assert_debug</a> (min (<a class="code" href="classbdm_1_1euni.html#ad3c27e331f90c754d80228108de8ed4c" title="internal">distance</a>) &gt; 0.0, <span class="stringliteral">&quot;bad support&quot;</span>); 
     409<a name="l00493"></a>00493                         <a class="code" href="classbdm_1_1euni.html#aff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a> = low0; 
     410<a name="l00494"></a>00494                         <a class="code" href="classbdm_1_1euni.html#acfad2dea4a62db6872bda8abd75f0de1" title="upper bound on support">high</a> = high0; 
     411<a name="l00495"></a>00495                         <a class="code" href="classbdm_1_1euni.html#a31bb13e8449a8eff35246d46dae35c20" title="normalizing coefficients">nk</a> = prod (1.0 / <a class="code" href="classbdm_1_1euni.html#ad3c27e331f90c754d80228108de8ed4c" title="internal">distance</a>); 
     412<a name="l00496"></a>00496                         <a class="code" href="classbdm_1_1euni.html#a3e63be48dd58659663ca60cd18700476" title="cache of log( nk )">lnk</a> = log (<a class="code" href="classbdm_1_1euni.html#a31bb13e8449a8eff35246d46dae35c20" title="normalizing coefficients">nk</a>); 
     413<a name="l00497"></a>00497                         <a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> = <a class="code" href="classbdm_1_1euni.html#aff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a>.length(); 
     414<a name="l00498"></a>00498                 } 
     415<a name="l00500"></a>00500  
     416<a name="l00501"></a><a class="code" href="classbdm_1_1euni.html#acaa07b8307bd793d5339d6583e0aba81">00501</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1euni.html#acaa07b8307bd793d5339d6583e0aba81">evallog</a> (<span class="keyword">const</span> vec &amp;val)<span class="keyword"> const  </span>{ 
     417<a name="l00502"></a>00502                         <span class="keywordflow">if</span> (any (val &lt; <a class="code" href="classbdm_1_1euni.html#aff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a>) &amp;&amp; any (val &gt; <a class="code" href="classbdm_1_1euni.html#acfad2dea4a62db6872bda8abd75f0de1" title="upper bound on support">high</a>)) {<span class="keywordflow">return</span> inf;} 
     418<a name="l00503"></a>00503                         <span class="keywordflow">else</span> <span class="keywordflow">return</span> <a class="code" href="classbdm_1_1euni.html#a3e63be48dd58659663ca60cd18700476" title="cache of log( nk )">lnk</a>; 
     419<a name="l00504"></a>00504                 } 
     420<a name="l00505"></a><a class="code" href="classbdm_1_1euni.html#afc5df80359ead2918384b2004ce67194">00505</a>                 vec <a class="code" href="classbdm_1_1euni.html#afc5df80359ead2918384b2004ce67194" title="Returns a sample,  from density .">sample</a>()<span class="keyword"> const </span>{ 
     421<a name="l00506"></a>00506                         vec smp (<a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>); 
     422<a name="l00507"></a>00507 <span class="preprocessor">#pragma omp critical</span> 
     423<a name="l00508"></a>00508 <span class="preprocessor"></span>                        UniRNG.sample_vector (<a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> , smp); 
     424<a name="l00509"></a>00509                         <span class="keywordflow">return</span> <a class="code" href="classbdm_1_1euni.html#aff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a> + elem_mult (<a class="code" href="classbdm_1_1euni.html#ad3c27e331f90c754d80228108de8ed4c" title="internal">distance</a>, smp); 
     425<a name="l00510"></a>00510                 } 
     426<a name="l00512"></a><a class="code" href="classbdm_1_1euni.html#a46caa8c13aba2e6228f964208918b226">00512</a>                 vec <a class="code" href="classbdm_1_1euni.html#a46caa8c13aba2e6228f964208918b226" title="set values of low and high ">mean</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> (<a class="code" href="classbdm_1_1euni.html#acfad2dea4a62db6872bda8abd75f0de1" title="upper bound on support">high</a> -<a class="code" href="classbdm_1_1euni.html#aff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a>) / 2.0;} 
     427<a name="l00513"></a><a class="code" href="classbdm_1_1euni.html#a951f932155111f6053c980f672b4c22c">00513</a>                 vec <a class="code" href="classbdm_1_1euni.html#a951f932155111f6053c980f672b4c22c" title="return expected variance (not covariance!)">variance</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> (pow (<a class="code" href="classbdm_1_1euni.html#acfad2dea4a62db6872bda8abd75f0de1" title="upper bound on support">high</a>, 2) + pow (<a class="code" href="classbdm_1_1euni.html#aff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a>, 2) + elem_mult (<a class="code" href="classbdm_1_1euni.html#acfad2dea4a62db6872bda8abd75f0de1" title="upper bound on support">high</a>, <a class="code" href="classbdm_1_1euni.html#aff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a>)) / 3.0;} 
     428<a name="l00522"></a><a class="code" href="classbdm_1_1euni.html#a77f5fef1f006fe056066da23b9e5f042">00522</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1euni.html#a77f5fef1f006fe056066da23b9e5f042">from_setting</a> (<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>) { 
     429<a name="l00523"></a>00523                         <a class="code" href="classbdm_1_1euni.html#a77f5fef1f006fe056066da23b9e5f042">epdf::from_setting</a> (<span class="keyword">set</span>); <span class="comment">// reads rv and rvc</span> 
    422430<a name="l00524"></a>00524  
    423 <a name="l00525"></a>00525  
    424 <a name="l00531"></a>00531         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    425 <a name="l00532"></a><a class="code" href="classbdm_1_1mlnorm.html">00532</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1mlnorm.html" title="Normal distributed linear function with linear function of mean value;.">mlnorm</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mEF.html" title="Exponential family model.">mEF</a> 
    426 <a name="l00533"></a>00533         { 
    427 <a name="l00534"></a>00534                 <span class="keyword">protected</span>: 
    428 <a name="l00536"></a><a class="code" href="classbdm_1_1mlnorm.html#6de45ab731143e70a94318df586c36ba">00536</a>                         <a class="code" href="classbdm_1_1shared__ptr.html" title="A naive implementation of roughly a subset of the std::tr1:shared_ptr spec (really...">shared_ptr&lt;enorm&lt;sq_T&gt;</a> &gt; <a class="code" href="classbdm_1_1mlnorm.html#6de45ab731143e70a94318df586c36ba" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>; 
    429 <a name="l00537"></a>00537                         mat A; 
    430 <a name="l00538"></a>00538                         vec mu_const; 
    431 <a name="l00539"></a>00539                         vec&amp; _mu; <span class="comment">//cached epdf.mu;</span> 
    432 <a name="l00540"></a>00540                 <span class="keyword">public</span>: 
    433 <a name="l00543"></a>00543                         <a class="code" href="classbdm_1_1mlnorm.html" title="Normal distributed linear function with linear function of mean value;.">mlnorm</a>():<a class="code" href="classbdm_1_1mlnorm.html#6de45ab731143e70a94318df586c36ba" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>(new <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm</a>&lt;sq_T&gt;()), _mu(<a class="code" href="classbdm_1_1mlnorm.html#6de45ab731143e70a94318df586c36ba" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>-&gt;_mu()) { set_ep(<a class="code" href="classbdm_1_1mlnorm.html#6de45ab731143e70a94318df586c36ba" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>); }; 
    434 <a name="l00544"></a>00544                         <a class="code" href="classbdm_1_1mlnorm.html" title="Normal distributed linear function with linear function of mean value;.">mlnorm</a> (<span class="keyword">const</span> mat &amp;A, <span class="keyword">const</span> vec &amp;mu0, <span class="keyword">const</span> sq_T &amp;R ) :<a class="code" href="classbdm_1_1mlnorm.html#6de45ab731143e70a94318df586c36ba" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>(new <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm</a>&lt;sq_T&gt;()), _mu(<a class="code" href="classbdm_1_1mlnorm.html#6de45ab731143e70a94318df586c36ba" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>-&gt;_mu()) 
    435 <a name="l00545"></a>00545                         { 
    436 <a name="l00546"></a>00546                                 set_ep(<a class="code" href="classbdm_1_1mlnorm.html#6de45ab731143e70a94318df586c36ba" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>); <a class="code" href="classbdm_1_1mlnorm.html#5d18dec3167584338a4775c1d165d96f" title="Set A and R.">set_parameters</a> ( A,mu0,R ); 
    437 <a name="l00547"></a>00547                         } 
    438 <a name="l00548"></a>00548  
    439 <a name="l00550"></a>00550                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlnorm.html#5d18dec3167584338a4775c1d165d96f" title="Set A and R.">set_parameters</a> ( <span class="keyword">const</span>  mat &amp;A, <span class="keyword">const</span> vec &amp;mu0, <span class="keyword">const</span> sq_T &amp;R ); 
    440 <a name="l00553"></a>00553                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlnorm.html#0dafc0196e7e07fd06dc6716e0e18bc2">condition</a> ( <span class="keyword">const</span> vec &amp;cond ); 
    441 <a name="l00554"></a>00554  
    442 <a name="l00556"></a><a class="code" href="classbdm_1_1mlnorm.html#56e587952f94fcac6cfc999eae6dbced">00556</a>                         vec&amp; <a class="code" href="classbdm_1_1mlnorm.html#56e587952f94fcac6cfc999eae6dbced" title="access function">_mu_const</a>() {<span class="keywordflow">return</span> mu_const;} 
    443 <a name="l00558"></a><a class="code" href="classbdm_1_1mlnorm.html#262a2a486bff585f34bb6a5005b02614">00558</a>                         mat&amp; <a class="code" href="classbdm_1_1mlnorm.html#262a2a486bff585f34bb6a5005b02614" title="access function">_A</a>() {<span class="keywordflow">return</span> A;} 
    444 <a name="l00560"></a><a class="code" href="classbdm_1_1mlnorm.html#78120ecd1c2b1d7e80124b4603504604">00560</a>                         mat <a class="code" href="classbdm_1_1mlnorm.html#78120ecd1c2b1d7e80124b4603504604" title="access function">_R</a>() { <span class="keywordflow">return</span> <a class="code" href="classbdm_1_1mlnorm.html#6de45ab731143e70a94318df586c36ba" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>-&gt;_R().to_mat(); } 
    445 <a name="l00561"></a>00561  
    446 <a name="l00562"></a>00562                         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_M&gt; 
    447 <a name="l00563"></a>00563                         <span class="keyword">friend</span> std::ostream &amp;operator&lt;&lt; ( std::ostream &amp;os,  mlnorm&lt;sq_M&gt; &amp;ml ); 
    448 <a name="l00564"></a>00564                          
    449 <a name="l00565"></a><a class="code" href="classbdm_1_1mlnorm.html#8582928555c309a2865f385dd0441fd1">00565</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlnorm.html#8582928555c309a2865f385dd0441fd1">from_setting</a>(<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>){ 
    450 <a name="l00566"></a>00566                                 <a class="code" href="classbdm_1_1mlnorm.html#8582928555c309a2865f385dd0441fd1">mpdf::from_setting</a>(<span class="keyword">set</span>);         
    451 <a name="l00567"></a>00567  
    452 <a name="l00568"></a>00568                                 <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>(A,<span class="keyword">set</span>,<span class="stringliteral">"A"</span>, UI::compulsory); 
    453 <a name="l00569"></a>00569                                 <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>(mu_const,<span class="keyword">set</span>,<span class="stringliteral">"const"</span>, UI::compulsory); 
    454 <a name="l00570"></a>00570                                 mat R0; 
    455 <a name="l00571"></a>00571                                 <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>(R0,<span class="keyword">set</span>,<span class="stringliteral">"R"</span>, UI::compulsory); 
    456 <a name="l00572"></a>00572                                 <a class="code" href="classbdm_1_1mlnorm.html#5d18dec3167584338a4775c1d165d96f" title="Set A and R.">set_parameters</a>(A,mu_const,R0); 
    457 <a name="l00573"></a>00573                         }; 
    458 <a name="l00574"></a>00574         }; 
    459 <a name="l00575"></a>00575         <a class="code" href="user__info_8h.html#4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a>(mlnorm&lt;ldmat&gt;); 
    460 <a name="l00576"></a>00576         <a class="code" href="user__info_8h.html#4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a>(mlnorm&lt;fsqmat&gt;); 
    461 <a name="l00577"></a>00577         <a class="code" href="user__info_8h.html#4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a>(mlnorm&lt;chmat&gt;); 
    462 <a name="l00578"></a>00578  
    463 <a name="l00580"></a>00580         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    464 <a name="l00581"></a><a class="code" href="classbdm_1_1mgnorm.html">00581</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1mgnorm.html" title="Mpdf with general function for mean value.">mgnorm</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mEF.html" title="Exponential family model.">mEF</a> 
    465 <a name="l00582"></a>00582         { 
    466 <a name="l00583"></a>00583                 <span class="keyword">protected</span>: 
    467 <a name="l00585"></a><a class="code" href="classbdm_1_1mgnorm.html#e63ba612a75f761c03025d22463219cb">00585</a>                         <a class="code" href="classbdm_1_1shared__ptr.html" title="A naive implementation of roughly a subset of the std::tr1:shared_ptr spec (really...">shared_ptr&lt;enorm&lt;sq_T&gt;</a> &gt; <a class="code" href="classbdm_1_1mgnorm.html#e63ba612a75f761c03025d22463219cb" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>; 
    468 <a name="l00586"></a>00586                         vec &amp;mu; 
    469 <a name="l00587"></a>00587                         <a class="code" href="classbdm_1_1fnc.html" title="Class representing function  of variable  represented by rv.">fnc</a>* g; 
    470 <a name="l00588"></a>00588                 <span class="keyword">public</span>: 
    471 <a name="l00590"></a><a class="code" href="classbdm_1_1mgnorm.html#1b014915d74470d3efab74e07cacb97d">00590</a>                         <a class="code" href="classbdm_1_1mgnorm.html#1b014915d74470d3efab74e07cacb97d" title="default constructor">mgnorm</a>():<a class="code" href="classbdm_1_1mgnorm.html#e63ba612a75f761c03025d22463219cb" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>(new <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm</a>&lt;sq_T&gt;()), mu(<a class="code" href="classbdm_1_1mgnorm.html#e63ba612a75f761c03025d22463219cb" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>-&gt;_mu()) { set_ep(<a class="code" href="classbdm_1_1mgnorm.html#e63ba612a75f761c03025d22463219cb" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>); } 
    472 <a name="l00592"></a><a class="code" href="classbdm_1_1mgnorm.html#578e02458e2a0d17f3864826b6ebd564">00592</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgnorm.html#578e02458e2a0d17f3864826b6ebd564" title="set mean function">set_parameters</a> ( <a class="code" href="classbdm_1_1fnc.html" title="Class representing function  of variable  represented by rv.">fnc</a>* g0, <span class="keyword">const</span> sq_T &amp;R0 ) {g=g0; <a class="code" href="classbdm_1_1mgnorm.html#e63ba612a75f761c03025d22463219cb" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>-&gt;set_parameters ( zeros ( g-&gt;<a class="code" href="classbdm_1_1fnc.html#083832294da9d1e40804158b979c4341" title="access function">dimension</a>() ), R0 );} 
    473 <a name="l00593"></a><a class="code" href="classbdm_1_1mgnorm.html#b31d63472cf6a1030cd8dbd8094c1f6d">00593</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgnorm.html#b31d63472cf6a1030cd8dbd8094c1f6d" title="Update ep so that it represents this mpdf conditioned on rvc = cond.">condition</a> ( <span class="keyword">const</span> vec &amp;cond ) {mu=g-&gt;<a class="code" href="classbdm_1_1fnc.html#6277b11d7fffc7ef8a2fa3e84ae5bad4" title="function evaluates numerical value of  at  cond ">eval</a> ( cond );}; 
    474 <a name="l00594"></a>00594  
    475 <a name="l00595"></a>00595  
    476 <a name="l00623"></a><a class="code" href="classbdm_1_1mgnorm.html#d717dacc6a9eb967f8410994dc6dc6f9">00623</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgnorm.html#d717dacc6a9eb967f8410994dc6dc6f9">from_setting</a>( <span class="keyword">const</span> Setting &amp;<span class="keyword">set</span> )  
    477 <a name="l00624"></a>00624                         {        
    478 <a name="l00625"></a>00625                                 <a class="code" href="classbdm_1_1fnc.html" title="Class representing function  of variable  represented by rv.">fnc</a>* g = UI::build&lt;fnc&gt;( <span class="keyword">set</span>, <span class="stringliteral">"g"</span>, UI::compulsory ); 
    479 <a name="l00626"></a>00626  
    480 <a name="l00627"></a>00627                                 mat R;                                   
    481 <a name="l00628"></a>00628                                 vec dR;                                  
    482 <a name="l00629"></a>00629                                 <span class="keywordflow">if</span> ( <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>( dR, <span class="keyword">set</span>, <span class="stringliteral">"dR"</span> ) ) 
    483 <a name="l00630"></a>00630                                         R=diag(dR); 
    484 <a name="l00631"></a>00631                                 <span class="keywordflow">else</span>  
    485 <a name="l00632"></a>00632                                         <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>( R, <span class="keyword">set</span>, <span class="stringliteral">"R"</span>, UI::compulsory);  
    486 <a name="l00633"></a>00633                  
    487 <a name="l00634"></a>00634                                 <a class="code" href="classbdm_1_1mgnorm.html#578e02458e2a0d17f3864826b6ebd564" title="set mean function">set_parameters</a>(g,R); 
    488 <a name="l00635"></a>00635                         } 
    489 <a name="l00636"></a>00636         }; 
    490 <a name="l00637"></a>00637  
    491 <a name="l00638"></a>00638         <a class="code" href="user__info_8h.html#4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a>(mgnorm&lt;chmat&gt;); 
    492 <a name="l00639"></a>00639  
    493 <a name="l00640"></a>00640  
    494 <a name="l00648"></a><a class="code" href="classbdm_1_1mlstudent.html">00648</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1mlstudent.html">mlstudent</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mlnorm.html" title="Normal distributed linear function with linear function of mean value;.">mlnorm</a>&lt;ldmat&gt; 
    495 <a name="l00649"></a>00649         { 
    496 <a name="l00650"></a>00650                 <span class="keyword">protected</span>: 
    497 <a name="l00651"></a>00651                         <a class="code" href="classldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a> Lambda; 
    498 <a name="l00652"></a>00652                         <a class="code" href="classldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a> &amp;<a class="code" href="classbdm_1_1mlnorm.html#78120ecd1c2b1d7e80124b4603504604" title="access function">_R</a>; 
    499 <a name="l00653"></a>00653                         <a class="code" href="classldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a> Re; 
    500 <a name="l00654"></a>00654                 <span class="keyword">public</span>: 
    501 <a name="l00655"></a>00655                         <a class="code" href="classbdm_1_1mlstudent.html">mlstudent</a> ( ) :<a class="code" href="classbdm_1_1mlnorm.html">mlnorm&lt;ldmat&gt;</a> (), 
    502 <a name="l00656"></a>00656                                         Lambda (),      <a class="code" href="classbdm_1_1mlnorm.html#78120ecd1c2b1d7e80124b4603504604" title="access function">_R</a> ( <a class="code" href="classbdm_1_1mlnorm.html#6de45ab731143e70a94318df586c36ba" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>-&gt;_R() ) {} 
    503 <a name="l00657"></a>00657                         <span class="keywordtype">void</span> set_parameters ( <span class="keyword">const</span> mat &amp;A0, <span class="keyword">const</span> vec &amp;mu0, <span class="keyword">const</span> <a class="code" href="classldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a> &amp;R0, <span class="keyword">const</span> <a class="code" href="classldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a>&amp; Lambda0 ) 
    504 <a name="l00658"></a>00658                         { 
    505 <a name="l00659"></a>00659                                 it_assert_debug ( A0.rows() ==mu0.length(),<span class="stringliteral">""</span> ); 
    506 <a name="l00660"></a>00660                                 it_assert_debug ( R0.<a class="code" href="classsqmat.html#071e80ced9cc3b8cbb360fa7462eb646" title="Reimplementing common functions of mat: rows().">rows</a>() ==A0.rows(),<span class="stringliteral">""</span> ); 
    507 <a name="l00661"></a>00661  
    508 <a name="l00662"></a>00662                                 <a class="code" href="classbdm_1_1mlnorm.html#6de45ab731143e70a94318df586c36ba" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>-&gt;set_parameters ( mu0,Lambda ); <span class="comment">//</span> 
    509 <a name="l00663"></a>00663                                 A = A0; 
    510 <a name="l00664"></a>00664                                 mu_const = mu0; 
    511 <a name="l00665"></a>00665                                 Re=R0; 
    512 <a name="l00666"></a>00666                                 Lambda = Lambda0; 
    513 <a name="l00667"></a>00667                         } 
    514 <a name="l00668"></a><a class="code" href="classbdm_1_1mlstudent.html#efd37560585c8613897f30d3c2f58d0d">00668</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlstudent.html#efd37560585c8613897f30d3c2f58d0d">condition</a> ( <span class="keyword">const</span> vec &amp;cond ) 
    515 <a name="l00669"></a>00669                         { 
    516 <a name="l00670"></a>00670                                 _mu = A*cond + mu_const; 
    517 <a name="l00671"></a>00671                                 <span class="keywordtype">double</span> zeta; 
    518 <a name="l00672"></a>00672                                 <span class="comment">//ugly hack!</span> 
    519 <a name="l00673"></a>00673                                 <span class="keywordflow">if</span> ( ( cond.length() +1 ) ==Lambda.<a class="code" href="classsqmat.html#071e80ced9cc3b8cbb360fa7462eb646" title="Reimplementing common functions of mat: rows().">rows</a>() ) 
    520 <a name="l00674"></a>00674                                 { 
    521 <a name="l00675"></a>00675                                         zeta = Lambda.<a class="code" href="classldmat.html#d876c5f83e02b3e809b35c9de5068f14" title="Evaluates quadratic form ;.">invqform</a> ( concat ( cond, vec_1 ( 1.0 ) ) ); 
    522 <a name="l00676"></a>00676                                 } 
    523 <a name="l00677"></a>00677                                 <span class="keywordflow">else</span> 
    524 <a name="l00678"></a>00678                                 { 
    525 <a name="l00679"></a>00679                                         zeta = Lambda.<a class="code" href="classldmat.html#d876c5f83e02b3e809b35c9de5068f14" title="Evaluates quadratic form ;.">invqform</a> ( cond ); 
    526 <a name="l00680"></a>00680                                 } 
    527 <a name="l00681"></a>00681                                 <a class="code" href="classbdm_1_1mlnorm.html#78120ecd1c2b1d7e80124b4603504604" title="access function">_R</a> = Re; 
    528 <a name="l00682"></a>00682                                 <a class="code" href="classbdm_1_1mlnorm.html#78120ecd1c2b1d7e80124b4603504604" title="access function">_R</a>*= ( 1+zeta );<span class="comment">// / ( nu ); &lt;&lt; nu is in Re!!!!!!</span> 
    529 <a name="l00683"></a>00683                         }; 
    530 <a name="l00684"></a>00684  
    531 <a name="l00685"></a>00685         }; 
    532 <a name="l00695"></a><a class="code" href="classbdm_1_1mgamma.html">00695</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1mgamma.html" title="Gamma random walk.">mgamma</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mEF.html" title="Exponential family model.">mEF</a> 
    533 <a name="l00696"></a>00696         { 
    534 <a name="l00697"></a>00697                 <span class="keyword">protected</span>: 
    535 <a name="l00699"></a><a class="code" href="classbdm_1_1mgamma.html#413aa1c71b58a71d64c08da3aa642c08">00699</a>                         <a class="code" href="classbdm_1_1shared__ptr.html">shared_ptr&lt;egamma&gt;</a> <a class="code" href="classbdm_1_1mgamma.html#413aa1c71b58a71d64c08da3aa642c08" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>; 
    536 <a name="l00700"></a>00700  
    537 <a name="l00702"></a><a class="code" href="classbdm_1_1mgamma.html#b20cf88cca1fe9b0b8f2a412608bfd09">00702</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1mgamma.html#b20cf88cca1fe9b0b8f2a412608bfd09" title="Constant .">k</a>; 
    538 <a name="l00703"></a>00703  
    539 <a name="l00705"></a><a class="code" href="classbdm_1_1mgamma.html#3d95f4dde9214ff6dba265e18af60312">00705</a>                         vec &amp;<a class="code" href="classbdm_1_1mgamma.html#3d95f4dde9214ff6dba265e18af60312" title="cache of iepdf.beta">_beta</a>; 
    540 <a name="l00706"></a>00706  
    541 <a name="l00707"></a>00707                 <span class="keyword">public</span>: 
    542 <a name="l00709"></a><a class="code" href="classbdm_1_1mgamma.html#1a9dc8661e5b214a8185d6e6b9956eb1">00709</a>                         <a class="code" href="classbdm_1_1mgamma.html#1a9dc8661e5b214a8185d6e6b9956eb1" title="Constructor.">mgamma</a>():<a class="code" href="classbdm_1_1mgamma.html#413aa1c71b58a71d64c08da3aa642c08" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>(new <a class="code" href="classbdm_1_1egamma.html" title="Gamma posterior density.">egamma</a>()), <a class="code" href="classbdm_1_1mgamma.html#b20cf88cca1fe9b0b8f2a412608bfd09" title="Constant .">k</a>(0), 
    543 <a name="l00710"></a>00710                             <a class="code" href="classbdm_1_1mgamma.html#3d95f4dde9214ff6dba265e18af60312" title="cache of iepdf.beta">_beta</a>(<a class="code" href="classbdm_1_1mgamma.html#413aa1c71b58a71d64c08da3aa642c08" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>-&gt;<a class="code" href="classbdm_1_1mgamma.html#3d95f4dde9214ff6dba265e18af60312" title="cache of iepdf.beta">_beta</a>()) { 
    544 <a name="l00711"></a>00711                             set_ep(<a class="code" href="classbdm_1_1mgamma.html#413aa1c71b58a71d64c08da3aa642c08" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>); 
    545 <a name="l00712"></a>00712                         } 
    546 <a name="l00713"></a>00713  
    547 <a name="l00715"></a>00715                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgamma.html#a0f21c2557b233a85838b497d040ab14" title="Set value of k.">set_parameters</a>(<span class="keywordtype">double</span> <a class="code" href="classbdm_1_1mgamma.html#b20cf88cca1fe9b0b8f2a412608bfd09" title="Constant .">k</a>, <span class="keyword">const</span> vec &amp;beta0); 
    548 <a name="l00716"></a>00716  
    549 <a name="l00717"></a><a class="code" href="classbdm_1_1mgamma.html#8996500f1885e39cde30221b20900bff">00717</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgamma.html#8996500f1885e39cde30221b20900bff" title="Update ep so that it represents this mpdf conditioned on rvc = cond.">condition</a> ( <span class="keyword">const</span> vec &amp;val ) {<a class="code" href="classbdm_1_1mgamma.html#3d95f4dde9214ff6dba265e18af60312" title="cache of iepdf.beta">_beta</a>=<a class="code" href="classbdm_1_1mgamma.html#b20cf88cca1fe9b0b8f2a412608bfd09" title="Constant .">k</a>/val;}; 
    550 <a name="l00727"></a><a class="code" href="classbdm_1_1mgamma.html#da2af0f327e5452bee71d1bf97452ae4">00727</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgamma.html#da2af0f327e5452bee71d1bf97452ae4">from_setting</a>(<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>){ 
    551 <a name="l00728"></a>00728                                 <a class="code" href="classbdm_1_1mgamma.html#da2af0f327e5452bee71d1bf97452ae4">mpdf::from_setting</a>(<span class="keyword">set</span>); <span class="comment">// reads rv and rvc</span> 
    552 <a name="l00729"></a>00729                                 vec betatmp; <span class="comment">// ugly but necessary</span> 
    553 <a name="l00730"></a>00730                                 <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>(betatmp,<span class="keyword">set</span>,<span class="stringliteral">"beta"</span>, UI::compulsory); 
    554 <a name="l00731"></a>00731                                 <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>(<a class="code" href="classbdm_1_1mgamma.html#b20cf88cca1fe9b0b8f2a412608bfd09" title="Constant .">k</a>,<span class="keyword">set</span>,<span class="stringliteral">"k"</span>, UI::compulsory); 
    555 <a name="l00732"></a>00732                                 <a class="code" href="classbdm_1_1mgamma.html#a0f21c2557b233a85838b497d040ab14" title="Set value of k.">set_parameters</a>(<a class="code" href="classbdm_1_1mgamma.html#b20cf88cca1fe9b0b8f2a412608bfd09" title="Constant .">k</a>,betatmp); 
    556 <a name="l00733"></a>00733                         } 
    557 <a name="l00734"></a>00734         }; 
    558 <a name="l00735"></a>00735         <a class="code" href="user__info_8h.html#4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a>(mgamma); 
    559 <a name="l00736"></a>00736          
    560 <a name="l00746"></a><a class="code" href="classbdm_1_1migamma.html">00746</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1migamma.html" title="Inverse-Gamma random walk.">migamma</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mEF.html" title="Exponential family model.">mEF</a> 
    561 <a name="l00747"></a>00747         { 
    562 <a name="l00748"></a>00748                 <span class="keyword">protected</span>: 
    563 <a name="l00750"></a><a class="code" href="classbdm_1_1migamma.html#f1636dbded763fecbd1c6a22dc6971fa">00750</a>                         <a class="code" href="classbdm_1_1shared__ptr.html">shared_ptr&lt;eigamma&gt;</a> <a class="code" href="classbdm_1_1migamma.html#f1636dbded763fecbd1c6a22dc6971fa" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>; 
     431<a name="l00525"></a>00525                         <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (<a class="code" href="classbdm_1_1euni.html#acfad2dea4a62db6872bda8abd75f0de1" title="upper bound on support">high</a>, <span class="keyword">set</span>, <span class="stringliteral">&quot;high&quot;</span>, UI::compulsory); 
     432<a name="l00526"></a>00526                         <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (<a class="code" href="classbdm_1_1euni.html#aff7ce6a2ef5ef0015bbd1398bed12f32" title="lower bound on support">low</a>, <span class="keyword">set</span>, <span class="stringliteral">&quot;low&quot;</span>, UI::compulsory); 
     433<a name="l00527"></a>00527                 } 
     434<a name="l00528"></a>00528 }; 
     435<a name="l00529"></a>00529  
     436<a name="l00530"></a>00530  
     437<a name="l00536"></a>00536 <span class="keyword">template</span> &lt; <span class="keyword">class</span> sq_T, <span class="keyword">template</span> &lt;<span class="keyword">typename</span>&gt; <span class="keyword">class </span>TEpdf = enorm &gt; 
     438<a name="l00537"></a><a class="code" href="classbdm_1_1mlnorm.html">00537</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1mlnorm.html" title="Normal distributed linear function with linear function of mean value;.">mlnorm</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mpdf__internal.html" title="Mpdf with internal epdf that is modified by function condition.">mpdf_internal</a>&lt; TEpdf&lt;sq_T&gt; &gt; 
     439<a name="l00538"></a>00538 { 
     440<a name="l00539"></a>00539         <span class="keyword">protected</span>: 
     441<a name="l00541"></a><a class="code" href="classbdm_1_1mlnorm.html#ac71bbf59fc732beb2b459d7117bd6e42">00541</a>                 mat <a class="code" href="classbdm_1_1mlnorm.html#ac71bbf59fc732beb2b459d7117bd6e42" title="Internal epdf that arise by conditioning on rvc.">A</a>; 
     442<a name="l00543"></a><a class="code" href="classbdm_1_1mlnorm.html#a831b50121cad10aa03a16682bd0c3ed6">00543</a>                 vec <a class="code" href="classbdm_1_1mlnorm.html#a831b50121cad10aa03a16682bd0c3ed6" title="Constant additive term.">mu_const</a>; 
     443<a name="l00544"></a>00544 <span class="comment">//                      vec&amp; _mu; //cached epdf.mu; !!!!!! WHY NOT?</span> 
     444<a name="l00545"></a>00545         <span class="keyword">public</span>: 
     445<a name="l00548"></a>00548                 <a class="code" href="classbdm_1_1mlnorm.html" title="Normal distributed linear function with linear function of mean value;.">mlnorm</a>() : <a class="code" href="classbdm_1_1mpdf__internal.html" title="Mpdf with internal epdf that is modified by function condition.">mpdf_internal</a>&lt; TEpdf&lt;sq_T&gt; &gt;() {}; 
     446<a name="l00549"></a>00549                 <a class="code" href="classbdm_1_1mlnorm.html" title="Normal distributed linear function with linear function of mean value;.">mlnorm</a> (<span class="keyword">const</span> mat &amp;<a class="code" href="classbdm_1_1mlnorm.html#ac71bbf59fc732beb2b459d7117bd6e42" title="Internal epdf that arise by conditioning on rvc.">A</a>, <span class="keyword">const</span> vec &amp;mu0, <span class="keyword">const</span> sq_T &amp;R) : <a class="code" href="classbdm_1_1mpdf__internal.html" title="Mpdf with internal epdf that is modified by function condition.">mpdf_internal</a>&lt; TEpdf&lt;sq_T&gt; &gt;() { 
     447<a name="l00550"></a>00550                         <a class="code" href="classbdm_1_1mlnorm.html#a04f7c6cda7b2f95161dd5fbcf15d1fd5" title="Set A and R.">set_parameters</a> (A, mu0, R); 
     448<a name="l00551"></a>00551                 } 
     449<a name="l00552"></a>00552  
     450<a name="l00554"></a><a class="code" href="classbdm_1_1mlnorm.html#a04f7c6cda7b2f95161dd5fbcf15d1fd5">00554</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlnorm.html#a04f7c6cda7b2f95161dd5fbcf15d1fd5" title="Set A and R.">set_parameters</a> (<span class="keyword">const</span>  mat &amp;A0, <span class="keyword">const</span> vec &amp;mu0, <span class="keyword">const</span> sq_T &amp;R0) { 
     451<a name="l00555"></a>00555                         <a class="code" href="bdmerror_8h.html#a89a0f906b242b79c5d3d342291b2cab4" title="Throw std::runtime_exception if t is not true and NDEBUG is not defined.">bdm_assert_debug</a> (A0.rows() == mu0.length(), <span class="stringliteral">&quot;mlnorm: A vs. mu mismatch&quot;</span>); 
     452<a name="l00556"></a>00556                         <a class="code" href="bdmerror_8h.html#a89a0f906b242b79c5d3d342291b2cab4" title="Throw std::runtime_exception if t is not true and NDEBUG is not defined.">bdm_assert_debug</a> (A0.rows() == R0.rows(), <span class="stringliteral">&quot;mlnorm: A vs. R mismatch&quot;</span>); 
     453<a name="l00557"></a>00557  
     454<a name="l00558"></a>00558                         this-&gt;<a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>.set_parameters (zeros (A0.rows()), R0); 
     455<a name="l00559"></a>00559                         <a class="code" href="classbdm_1_1mlnorm.html#ac71bbf59fc732beb2b459d7117bd6e42" title="Internal epdf that arise by conditioning on rvc.">A</a> = A0; 
     456<a name="l00560"></a>00560                         <a class="code" href="classbdm_1_1mlnorm.html#a831b50121cad10aa03a16682bd0c3ed6" title="Constant additive term.">mu_const</a> = mu0; 
     457<a name="l00561"></a>00561                         this-&gt;<a class="code" href="classbdm_1_1mpdf.html#a7c1900976ff13dbc09c9729b3bbff9e6" title="dimension of the condition">dimc</a> = A0.cols(); 
     458<a name="l00562"></a>00562                 } 
     459<a name="l00565"></a><a class="code" href="classbdm_1_1mlnorm.html#ac2895ae549ee76d961be98d7061bd110">00565</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlnorm.html#ac2895ae549ee76d961be98d7061bd110">condition</a> (<span class="keyword">const</span> vec &amp;cond) { 
     460<a name="l00566"></a>00566                         this-&gt;<a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>._mu() = <a class="code" href="classbdm_1_1mlnorm.html#ac71bbf59fc732beb2b459d7117bd6e42" title="Internal epdf that arise by conditioning on rvc.">A</a> * cond + <a class="code" href="classbdm_1_1mlnorm.html#a831b50121cad10aa03a16682bd0c3ed6" title="Constant additive term.">mu_const</a>; 
     461<a name="l00567"></a>00567 <span class="comment">//R is already assigned;</span> 
     462<a name="l00568"></a>00568                 } 
     463<a name="l00569"></a>00569  
     464<a name="l00571"></a><a class="code" href="classbdm_1_1mlnorm.html#a6332e5200f3afa15db3f7f4bca09b17f">00571</a>                 vec&amp; <a class="code" href="classbdm_1_1mlnorm.html#a6332e5200f3afa15db3f7f4bca09b17f" title="access function">_mu_const</a>() {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1mlnorm.html#a831b50121cad10aa03a16682bd0c3ed6" title="Constant additive term.">mu_const</a>;} 
     465<a name="l00573"></a><a class="code" href="classbdm_1_1mlnorm.html#ab256b547c5156b5898a3a1e5462f9540">00573</a>                 mat&amp; <a class="code" href="classbdm_1_1mlnorm.html#ab256b547c5156b5898a3a1e5462f9540" title="access function">_A</a>() {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1mlnorm.html#ac71bbf59fc732beb2b459d7117bd6e42" title="Internal epdf that arise by conditioning on rvc.">A</a>;} 
     466<a name="l00575"></a><a class="code" href="classbdm_1_1mlnorm.html#a564715dea56f7bbff0083aec295ce97e">00575</a>                 mat <a class="code" href="classbdm_1_1mlnorm.html#a564715dea56f7bbff0083aec295ce97e" title="access function">_R</a>() { <span class="keywordflow">return</span> this-&gt;<a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>._R().to_mat(); } 
     467<a name="l00576"></a>00576  
     468<a name="l00578"></a>00578                 <span class="keyword">template</span>&lt;<span class="keyword">typename</span> sq_M&gt; 
     469<a name="l00579"></a>00579                 <span class="keyword">friend</span> std::ostream &amp;operator&lt;&lt; (std::ostream &amp;os,  mlnorm&lt;sq_M, enorm&gt; &amp;ml); 
     470<a name="l00580"></a>00580  
     471<a name="l00581"></a><a class="code" href="classbdm_1_1mlnorm.html#a52980f13d80162d00b30d5864343f564">00581</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlnorm.html#a52980f13d80162d00b30d5864343f564">from_setting</a> (<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>) { 
     472<a name="l00582"></a>00582                         <a class="code" href="classbdm_1_1mlnorm.html#a52980f13d80162d00b30d5864343f564">mpdf::from_setting</a> (<span class="keyword">set</span>); 
     473<a name="l00583"></a>00583  
     474<a name="l00584"></a>00584                         <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (<a class="code" href="classbdm_1_1mlnorm.html#ac71bbf59fc732beb2b459d7117bd6e42" title="Internal epdf that arise by conditioning on rvc.">A</a>, <span class="keyword">set</span>, <span class="stringliteral">&quot;A&quot;</span>, UI::compulsory); 
     475<a name="l00585"></a>00585                         <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (<a class="code" href="classbdm_1_1mlnorm.html#a831b50121cad10aa03a16682bd0c3ed6" title="Constant additive term.">mu_const</a>, <span class="keyword">set</span>, <span class="stringliteral">&quot;const&quot;</span>, UI::compulsory); 
     476<a name="l00586"></a>00586                         mat R0; 
     477<a name="l00587"></a>00587                         <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (R0, <span class="keyword">set</span>, <span class="stringliteral">&quot;R&quot;</span>, UI::compulsory); 
     478<a name="l00588"></a>00588                         <a class="code" href="classbdm_1_1mlnorm.html#a04f7c6cda7b2f95161dd5fbcf15d1fd5" title="Set A and R.">set_parameters</a> (<a class="code" href="classbdm_1_1mlnorm.html#ac71bbf59fc732beb2b459d7117bd6e42" title="Internal epdf that arise by conditioning on rvc.">A</a>, <a class="code" href="classbdm_1_1mlnorm.html#a831b50121cad10aa03a16682bd0c3ed6" title="Constant additive term.">mu_const</a>, R0); 
     479<a name="l00589"></a>00589                 }; 
     480<a name="l00590"></a>00590 }; 
     481<a name="l00591"></a>00591 <a class="code" href="user__info_8h.html#ab06f1e7c44a8306e321f62a0099210b9" title="Variant of UIREGISTER for templated classes.">UIREGISTER2</a> (mlnorm,ldmat); 
     482<a name="l00592"></a>00592 SHAREDPTR2 ( mlnorm, ldmat ); 
     483<a name="l00593"></a>00593 <a class="code" href="user__info_8h.html#ab06f1e7c44a8306e321f62a0099210b9" title="Variant of UIREGISTER for templated classes.">UIREGISTER2</a> (mlnorm,fsqmat); 
     484<a name="l00594"></a>00594 SHAREDPTR2 ( mlnorm, fsqmat ); 
     485<a name="l00595"></a>00595 <a class="code" href="user__info_8h.html#ab06f1e7c44a8306e321f62a0099210b9" title="Variant of UIREGISTER for templated classes.">UIREGISTER2</a> (mlnorm, chmat); 
     486<a name="l00596"></a>00596 SHAREDPTR2 ( mlnorm, chmat ); 
     487<a name="l00597"></a>00597  
     488<a name="l00599"></a>00599 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     489<a name="l00600"></a><a class="code" href="classbdm_1_1mgnorm.html">00600</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1mgnorm.html" title="Mpdf with general function for mean value.">mgnorm</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mpdf__internal.html" title="Mpdf with internal epdf that is modified by function condition.">mpdf_internal</a>&lt; enorm&lt; sq_T &gt; &gt; 
     490<a name="l00601"></a>00601 { 
     491<a name="l00602"></a>00602         <span class="keyword">private</span>: 
     492<a name="l00603"></a>00603 <span class="comment">//                      vec &amp;mu; WHY NOT?</span> 
     493<a name="l00604"></a>00604                 <a class="code" href="classbdm_1_1shared__ptr.html">shared_ptr&lt;fnc&gt;</a> g; 
     494<a name="l00605"></a>00605  
     495<a name="l00606"></a>00606         <span class="keyword">public</span>: 
     496<a name="l00608"></a><a class="code" href="classbdm_1_1mgnorm.html#a1b014915d74470d3efab74e07cacb97d">00608</a>                 <a class="code" href="classbdm_1_1mgnorm.html#a1b014915d74470d3efab74e07cacb97d" title="default constructor">mgnorm</a>() : <a class="code" href="classbdm_1_1mpdf__internal.html" title="Mpdf with internal epdf that is modified by function condition.">mpdf_internal</a>&lt;<a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm</a>&lt;sq_T&gt; &gt;() { } 
     497<a name="l00610"></a>00610                 <span class="keyword">inline</span> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgnorm.html#ab736332d20e418bf50d45836e129f339" title="set mean function">set_parameters</a> (<span class="keyword">const</span> <a class="code" href="classbdm_1_1shared__ptr.html">shared_ptr&lt;fnc&gt;</a> &amp;g0, <span class="keyword">const</span> sq_T &amp;R0); 
     498<a name="l00611"></a>00611                 <span class="keyword">inline</span> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgnorm.html#ab31d63472cf6a1030cd8dbd8094c1f6d">condition</a> (<span class="keyword">const</span> vec &amp;cond); 
     499<a name="l00612"></a>00612  
     500<a name="l00613"></a>00613  
     501<a name="l00641"></a><a class="code" href="classbdm_1_1mgnorm.html#ad717dacc6a9eb967f8410994dc6dc6f9">00641</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgnorm.html#ad717dacc6a9eb967f8410994dc6dc6f9">from_setting</a> (<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>) { 
     502<a name="l00642"></a>00642                         <a class="code" href="classbdm_1_1shared__ptr.html">shared_ptr&lt;fnc&gt;</a> g = UI::build&lt;fnc&gt; (<span class="keyword">set</span>, <span class="stringliteral">&quot;g&quot;</span>, UI::compulsory); 
     503<a name="l00643"></a>00643  
     504<a name="l00644"></a>00644                         mat R; 
     505<a name="l00645"></a>00645                         vec dR; 
     506<a name="l00646"></a>00646                         <span class="keywordflow">if</span> (<a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (dR, <span class="keyword">set</span>, <span class="stringliteral">&quot;dR&quot;</span>)) 
     507<a name="l00647"></a>00647                                 R = diag (dR); 
     508<a name="l00648"></a>00648                         <span class="keywordflow">else</span> 
     509<a name="l00649"></a>00649                                 <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (R, <span class="keyword">set</span>, <span class="stringliteral">&quot;R&quot;</span>, UI::compulsory); 
     510<a name="l00650"></a>00650  
     511<a name="l00651"></a>00651                         <a class="code" href="classbdm_1_1mgnorm.html#ab736332d20e418bf50d45836e129f339" title="set mean function">set_parameters</a> (g, R); 
     512<a name="l00652"></a>00652                 } 
     513<a name="l00653"></a>00653 }; 
     514<a name="l00654"></a>00654  
     515<a name="l00655"></a>00655 <a class="code" href="user__info_8h.html#ab06f1e7c44a8306e321f62a0099210b9" title="Variant of UIREGISTER for templated classes.">UIREGISTER2</a> (mgnorm, chmat); 
     516<a name="l00656"></a>00656 SHAREDPTR2 ( mgnorm, chmat ); 
     517<a name="l00657"></a>00657  
     518<a name="l00658"></a>00658  
     519<a name="l00666"></a><a class="code" href="classbdm_1_1mlstudent.html">00666</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1mlstudent.html">mlstudent</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mlnorm.html" title="Normal distributed linear function with linear function of mean value;.">mlnorm</a>&lt;ldmat, enorm&gt; 
     520<a name="l00667"></a>00667 { 
     521<a name="l00668"></a>00668         <span class="keyword">protected</span>: 
     522<a name="l00670"></a><a class="code" href="classbdm_1_1mlstudent.html#a41595144a79594acbe288c6b59412657">00670</a>                 <a class="code" href="classbdm_1_1ldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a> <a class="code" href="classbdm_1_1mlstudent.html#a41595144a79594acbe288c6b59412657" title="Variable  from theory.">Lambda</a>; 
     523<a name="l00672"></a><a class="code" href="classbdm_1_1mlstudent.html#a72e9bda4d6684e07faafc4b2192daf39">00672</a>                 <a class="code" href="classbdm_1_1ldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a> &amp;<a class="code" href="classbdm_1_1mlnorm.html#a564715dea56f7bbff0083aec295ce97e" title="access function">_R</a>; 
     524<a name="l00674"></a><a class="code" href="classbdm_1_1mlstudent.html#a1c063ad6cb6e079ee11bc4128c2c9fe8">00674</a>                 <a class="code" href="classbdm_1_1ldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a> <a class="code" href="classbdm_1_1mlstudent.html#a1c063ad6cb6e079ee11bc4128c2c9fe8" title="Variable .">Re</a>; 
     525<a name="l00675"></a>00675         <span class="keyword">public</span>: 
     526<a name="l00676"></a>00676                 <a class="code" href="classbdm_1_1mlstudent.html">mlstudent</a> () : <a class="code" href="classbdm_1_1mlnorm.html" title="Normal distributed linear function with linear function of mean value;.">mlnorm</a>&lt;<a class="code" href="classbdm_1_1ldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a>, <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm</a>&gt; (), 
     527<a name="l00677"></a>00677                                 <a class="code" href="classbdm_1_1mlstudent.html#a41595144a79594acbe288c6b59412657" title="Variable  from theory.">Lambda</a> (),      <a class="code" href="classbdm_1_1mlnorm.html#a564715dea56f7bbff0083aec295ce97e" title="access function">_R</a> (<a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>.<a class="code" href="classbdm_1_1mlnorm.html#a564715dea56f7bbff0083aec295ce97e" title="access function">_R</a>()) {} 
     528<a name="l00679"></a><a class="code" href="classbdm_1_1mlstudent.html#a4cdf79aac1b2165c0290e73810a0e4a3">00679</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlstudent.html#a4cdf79aac1b2165c0290e73810a0e4a3" title="constructor function">set_parameters</a> (<span class="keyword">const</span> mat &amp;A0, <span class="keyword">const</span> vec &amp;mu0, <span class="keyword">const</span> <a class="code" href="classbdm_1_1ldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a> &amp;R0, <span class="keyword">const</span> <a class="code" href="classbdm_1_1ldmat.html" title="Matrix stored in LD form, (commonly known as UD).">ldmat</a>&amp; Lambda0) { 
     529<a name="l00680"></a>00680                         <a class="code" href="bdmerror_8h.html#a89a0f906b242b79c5d3d342291b2cab4" title="Throw std::runtime_exception if t is not true and NDEBUG is not defined.">bdm_assert_debug</a> (A0.rows() == mu0.length(), <span class="stringliteral">&quot;mlstudent: A vs. mu mismatch&quot;</span>); 
     530<a name="l00681"></a>00681                         <a class="code" href="bdmerror_8h.html#a89a0f906b242b79c5d3d342291b2cab4" title="Throw std::runtime_exception if t is not true and NDEBUG is not defined.">bdm_assert_debug</a> (R0.<a class="code" href="classbdm_1_1sqmat.html#a73e639221343dcce76c3305524d67590" title="Reimplementing common functions of mat: rows().">rows</a>() == A0.rows(), <span class="stringliteral">&quot;mlstudent: A vs. R mismatch&quot;</span>); 
     531<a name="l00682"></a>00682  
     532<a name="l00683"></a>00683                         <a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>.set_parameters (mu0, R0);<span class="comment">// was Lambda, why?</span> 
     533<a name="l00684"></a>00684                         <a class="code" href="classbdm_1_1mlnorm.html#ac71bbf59fc732beb2b459d7117bd6e42" title="Internal epdf that arise by conditioning on rvc.">A</a> = A0; 
     534<a name="l00685"></a>00685                         <a class="code" href="classbdm_1_1mlnorm.html#a831b50121cad10aa03a16682bd0c3ed6" title="Constant additive term.">mu_const</a> = mu0; 
     535<a name="l00686"></a>00686                         <a class="code" href="classbdm_1_1mlstudent.html#a1c063ad6cb6e079ee11bc4128c2c9fe8" title="Variable .">Re</a> = R0; 
     536<a name="l00687"></a>00687                         <a class="code" href="classbdm_1_1mlstudent.html#a41595144a79594acbe288c6b59412657" title="Variable  from theory.">Lambda</a> = Lambda0; 
     537<a name="l00688"></a>00688                 } 
     538<a name="l00689"></a><a class="code" href="classbdm_1_1mlstudent.html#aefd37560585c8613897f30d3c2f58d0d">00689</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlstudent.html#aefd37560585c8613897f30d3c2f58d0d">condition</a> (<span class="keyword">const</span> vec &amp;cond) { 
     539<a name="l00690"></a>00690                         <a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>._mu() = <a class="code" href="classbdm_1_1mlnorm.html#ac71bbf59fc732beb2b459d7117bd6e42" title="Internal epdf that arise by conditioning on rvc.">A</a> * cond + <a class="code" href="classbdm_1_1mlnorm.html#a831b50121cad10aa03a16682bd0c3ed6" title="Constant additive term.">mu_const</a>; 
     540<a name="l00691"></a>00691                         <span class="keywordtype">double</span> zeta; 
     541<a name="l00692"></a>00692                         <span class="comment">//ugly hack!</span> 
     542<a name="l00693"></a>00693                         <span class="keywordflow">if</span> ( (cond.length() + 1) == <a class="code" href="classbdm_1_1mlstudent.html#a41595144a79594acbe288c6b59412657" title="Variable  from theory.">Lambda</a>.<a class="code" href="classbdm_1_1sqmat.html#a73e639221343dcce76c3305524d67590" title="Reimplementing common functions of mat: rows().">rows</a>()) { 
     543<a name="l00694"></a>00694                                 zeta = <a class="code" href="classbdm_1_1mlstudent.html#a41595144a79594acbe288c6b59412657" title="Variable  from theory.">Lambda</a>.<a class="code" href="classbdm_1_1ldmat.html#af743de194aadb8515cf18226fadf365f" title="Evaluates quadratic form ;.">invqform</a> (concat (cond, vec_1 (1.0))); 
     544<a name="l00695"></a>00695                         } <span class="keywordflow">else</span> { 
     545<a name="l00696"></a>00696                                 zeta = <a class="code" href="classbdm_1_1mlstudent.html#a41595144a79594acbe288c6b59412657" title="Variable  from theory.">Lambda</a>.<a class="code" href="classbdm_1_1ldmat.html#af743de194aadb8515cf18226fadf365f" title="Evaluates quadratic form ;.">invqform</a> (cond); 
     546<a name="l00697"></a>00697                         } 
     547<a name="l00698"></a>00698                         <a class="code" href="classbdm_1_1mlnorm.html#a564715dea56f7bbff0083aec295ce97e" title="access function">_R</a> = <a class="code" href="classbdm_1_1mlstudent.html#a1c063ad6cb6e079ee11bc4128c2c9fe8" title="Variable .">Re</a>; 
     548<a name="l00699"></a>00699                         <a class="code" href="classbdm_1_1mlnorm.html#a564715dea56f7bbff0083aec295ce97e" title="access function">_R</a> *= (1 + zeta);<span class="comment">// / ( nu ); &lt;&lt; nu is in Re!!!!!!</span> 
     549<a name="l00700"></a>00700                 }; 
     550<a name="l00701"></a>00701  
     551<a name="l00702"></a>00702 }; 
     552<a name="l00712"></a><a class="code" href="classbdm_1_1mgamma.html">00712</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1mgamma.html" title="Gamma random walk.">mgamma</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mpdf__internal.html" title="Mpdf with internal epdf that is modified by function condition.">mpdf_internal</a>&lt;egamma&gt; 
     553<a name="l00713"></a>00713 { 
     554<a name="l00714"></a>00714         <span class="keyword">protected</span>: 
     555<a name="l00715"></a>00715  
     556<a name="l00717"></a><a class="code" href="classbdm_1_1mgamma.html#ab20cf88cca1fe9b0b8f2a412608bfd09">00717</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1mgamma.html#ab20cf88cca1fe9b0b8f2a412608bfd09" title="Constant .">k</a>; 
     557<a name="l00718"></a>00718  
     558<a name="l00720"></a><a class="code" href="classbdm_1_1mgamma.html#a3d95f4dde9214ff6dba265e18af60312">00720</a>                 vec &amp;<a class="code" href="classbdm_1_1mgamma.html#a3d95f4dde9214ff6dba265e18af60312" title="cache of iepdf.beta">_beta</a>; 
     559<a name="l00721"></a>00721  
     560<a name="l00722"></a>00722         <span class="keyword">public</span>: 
     561<a name="l00724"></a><a class="code" href="classbdm_1_1mgamma.html#a1a9dc8661e5b214a8185d6e6b9956eb1">00724</a>                 <a class="code" href="classbdm_1_1mgamma.html#a1a9dc8661e5b214a8185d6e6b9956eb1" title="Constructor.">mgamma</a>() : <a class="code" href="classbdm_1_1mpdf__internal.html" title="Mpdf with internal epdf that is modified by function condition.">mpdf_internal</a>&lt;<a class="code" href="classbdm_1_1egamma.html" title="Gamma posterior density.">egamma</a>&gt;(), <a class="code" href="classbdm_1_1mgamma.html#ab20cf88cca1fe9b0b8f2a412608bfd09" title="Constant .">k</a> (0), 
     562<a name="l00725"></a>00725                                 <a class="code" href="classbdm_1_1mgamma.html#a3d95f4dde9214ff6dba265e18af60312" title="cache of iepdf.beta">_beta</a> (<a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>.<a class="code" href="classbdm_1_1mgamma.html#a3d95f4dde9214ff6dba265e18af60312" title="cache of iepdf.beta">_beta</a>()) { 
     563<a name="l00726"></a>00726                 } 
     564<a name="l00727"></a>00727  
     565<a name="l00729"></a>00729                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgamma.html#aa0f21c2557b233a85838b497d040ab14" title="Set value of k.">set_parameters</a> (<span class="keywordtype">double</span> <a class="code" href="classbdm_1_1mgamma.html#ab20cf88cca1fe9b0b8f2a412608bfd09" title="Constant .">k</a>, <span class="keyword">const</span> vec &amp;beta0); 
     566<a name="l00730"></a>00730  
     567<a name="l00731"></a><a class="code" href="classbdm_1_1mgamma.html#a8996500f1885e39cde30221b20900bff">00731</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgamma.html#a8996500f1885e39cde30221b20900bff">condition</a> (<span class="keyword">const</span> vec &amp;val) {<a class="code" href="classbdm_1_1mgamma.html#a3d95f4dde9214ff6dba265e18af60312" title="cache of iepdf.beta">_beta</a> = <a class="code" href="classbdm_1_1mgamma.html#ab20cf88cca1fe9b0b8f2a412608bfd09" title="Constant .">k</a> / val;}; 
     568<a name="l00741"></a><a class="code" href="classbdm_1_1mgamma.html#ada2af0f327e5452bee71d1bf97452ae4">00741</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgamma.html#ada2af0f327e5452bee71d1bf97452ae4">from_setting</a> (<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>) { 
     569<a name="l00742"></a>00742                         <a class="code" href="classbdm_1_1mgamma.html#ada2af0f327e5452bee71d1bf97452ae4">mpdf::from_setting</a> (<span class="keyword">set</span>); <span class="comment">// reads rv and rvc</span> 
     570<a name="l00743"></a>00743                         vec betatmp; <span class="comment">// ugly but necessary</span> 
     571<a name="l00744"></a>00744                         <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (betatmp, <span class="keyword">set</span>, <span class="stringliteral">&quot;beta&quot;</span>, UI::compulsory); 
     572<a name="l00745"></a>00745                         <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (<a class="code" href="classbdm_1_1mgamma.html#ab20cf88cca1fe9b0b8f2a412608bfd09" title="Constant .">k</a>, <span class="keyword">set</span>, <span class="stringliteral">&quot;k&quot;</span>, UI::compulsory); 
     573<a name="l00746"></a>00746                         <a class="code" href="classbdm_1_1mgamma.html#aa0f21c2557b233a85838b497d040ab14" title="Set value of k.">set_parameters</a> (<a class="code" href="classbdm_1_1mgamma.html#ab20cf88cca1fe9b0b8f2a412608bfd09" title="Constant .">k</a>, betatmp); 
     574<a name="l00747"></a>00747                 } 
     575<a name="l00748"></a>00748 }; 
     576<a name="l00749"></a>00749 <a class="code" href="user__info_8h.html#a4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a> (mgamma); 
     577<a name="l00750"></a>00750 SHAREDPTR (mgamma); 
    564578<a name="l00751"></a>00751  
    565 <a name="l00753"></a><a class="code" href="classbdm_1_1migamma.html#dc56bc9da542e0103ec16b9be8e5e38c">00753</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1migamma.html#dc56bc9da542e0103ec16b9be8e5e38c" title="Constant .">k</a>; 
    566 <a name="l00754"></a>00754  
    567 <a name="l00756"></a><a class="code" href="classbdm_1_1migamma.html#c9847093da59a9ba0ebb68d2c592f5dc">00756</a>                         vec &amp;<a class="code" href="classbdm_1_1migamma.html#c9847093da59a9ba0ebb68d2c592f5dc" title="cache of iepdf.alpha">_alpha</a>; 
    568 <a name="l00757"></a>00757  
    569 <a name="l00759"></a><a class="code" href="classbdm_1_1migamma.html#0d854c047001b5465cf1ba21f52904b5">00759</a>                         vec &amp;<a class="code" href="classbdm_1_1migamma.html#0d854c047001b5465cf1ba21f52904b5" title="cache of iepdf.beta">_beta</a>; 
    570 <a name="l00760"></a>00760  
    571 <a name="l00761"></a>00761                 <span class="keyword">public</span>: 
    572 <a name="l00764"></a>00764                         <a class="code" href="classbdm_1_1migamma.html" title="Inverse-Gamma random walk.">migamma</a>():<a class="code" href="classbdm_1_1migamma.html#f1636dbded763fecbd1c6a22dc6971fa" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>(new <a class="code" href="classbdm_1_1eigamma.html" title="Inverse-Gamma posterior density.">eigamma</a>()), 
    573 <a name="l00765"></a>00765                             <a class="code" href="classbdm_1_1migamma.html#dc56bc9da542e0103ec16b9be8e5e38c" title="Constant .">k</a>(0), 
    574 <a name="l00766"></a>00766                             <a class="code" href="classbdm_1_1migamma.html#c9847093da59a9ba0ebb68d2c592f5dc" title="cache of iepdf.alpha">_alpha</a>(<a class="code" href="classbdm_1_1migamma.html#f1636dbded763fecbd1c6a22dc6971fa" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>-&gt;<a class="code" href="classbdm_1_1migamma.html#c9847093da59a9ba0ebb68d2c592f5dc" title="cache of iepdf.alpha">_alpha</a>()), 
    575 <a name="l00767"></a>00767                             <a class="code" href="classbdm_1_1migamma.html#0d854c047001b5465cf1ba21f52904b5" title="cache of iepdf.beta">_beta</a>(<a class="code" href="classbdm_1_1migamma.html#f1636dbded763fecbd1c6a22dc6971fa" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>-&gt;<a class="code" href="classbdm_1_1migamma.html#0d854c047001b5465cf1ba21f52904b5" title="cache of iepdf.beta">_beta</a>()) { 
    576 <a name="l00768"></a>00768                             set_ep(<a class="code" href="classbdm_1_1migamma.html#f1636dbded763fecbd1c6a22dc6971fa" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>); 
    577 <a name="l00769"></a>00769                         } 
    578 <a name="l00770"></a>00770  
    579 <a name="l00771"></a>00771                         <a class="code" href="classbdm_1_1migamma.html" title="Inverse-Gamma random walk.">migamma</a>(<span class="keyword">const</span> <a class="code" href="classbdm_1_1migamma.html" title="Inverse-Gamma random walk.">migamma</a> &amp;m):<a class="code" href="classbdm_1_1migamma.html#f1636dbded763fecbd1c6a22dc6971fa" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>(m.<a class="code" href="classbdm_1_1migamma.html#f1636dbded763fecbd1c6a22dc6971fa" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>), 
    580 <a name="l00772"></a>00772                             <a class="code" href="classbdm_1_1migamma.html#dc56bc9da542e0103ec16b9be8e5e38c" title="Constant .">k</a>(0), 
    581 <a name="l00773"></a>00773                             <a class="code" href="classbdm_1_1migamma.html#c9847093da59a9ba0ebb68d2c592f5dc" title="cache of iepdf.alpha">_alpha</a>(<a class="code" href="classbdm_1_1migamma.html#f1636dbded763fecbd1c6a22dc6971fa" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>-&gt;<a class="code" href="classbdm_1_1migamma.html#c9847093da59a9ba0ebb68d2c592f5dc" title="cache of iepdf.alpha">_alpha</a>()), 
    582 <a name="l00774"></a>00774                             <a class="code" href="classbdm_1_1migamma.html#0d854c047001b5465cf1ba21f52904b5" title="cache of iepdf.beta">_beta</a>(<a class="code" href="classbdm_1_1migamma.html#f1636dbded763fecbd1c6a22dc6971fa" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>-&gt;<a class="code" href="classbdm_1_1migamma.html#0d854c047001b5465cf1ba21f52904b5" title="cache of iepdf.beta">_beta</a>()) { 
    583 <a name="l00775"></a>00775                             set_ep(<a class="code" href="classbdm_1_1migamma.html#f1636dbded763fecbd1c6a22dc6971fa" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>); 
    584 <a name="l00776"></a>00776                         } 
    585 <a name="l00778"></a>00778  
    586 <a name="l00780"></a><a class="code" href="classbdm_1_1migamma.html#8b10ab922e2a7bae2fb6bb3efc7b6151">00780</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1migamma.html#8b10ab922e2a7bae2fb6bb3efc7b6151" title="Set value of k.">set_parameters</a> ( <span class="keywordtype">int</span> len, <span class="keywordtype">double</span> k0 ) 
    587 <a name="l00781"></a>00781                         { 
    588 <a name="l00782"></a>00782                                 <a class="code" href="classbdm_1_1migamma.html#dc56bc9da542e0103ec16b9be8e5e38c" title="Constant .">k</a>=k0; 
    589 <a name="l00783"></a>00783                                 <a class="code" href="classbdm_1_1migamma.html#f1636dbded763fecbd1c6a22dc6971fa" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>-&gt;set_parameters ( ( 1.0/ ( <a class="code" href="classbdm_1_1migamma.html#dc56bc9da542e0103ec16b9be8e5e38c" title="Constant .">k</a>*<a class="code" href="classbdm_1_1migamma.html#dc56bc9da542e0103ec16b9be8e5e38c" title="Constant .">k</a> ) +2.0 ) *ones ( len ) <span class="comment">/*alpha*/</span>, ones ( len ) <span class="comment">/*beta*/</span> ); 
    590 <a name="l00784"></a>00784                                 <a class="code" href="classbdm_1_1mpdf.html#7c1900976ff13dbc09c9729b3bbff9e6" title="dimension of the condition">dimc</a> = dimension(); 
    591 <a name="l00785"></a>00785                         }; 
    592 <a name="l00786"></a><a class="code" href="classbdm_1_1migamma.html#7a34b1e2e3aa2250d7c0ed7df1665b8c">00786</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1migamma.html#7a34b1e2e3aa2250d7c0ed7df1665b8c" title="Update ep so that it represents this mpdf conditioned on rvc = cond.">condition</a> ( <span class="keyword">const</span> vec &amp;val ) 
    593 <a name="l00787"></a>00787                         { 
    594 <a name="l00788"></a>00788                                 <a class="code" href="classbdm_1_1migamma.html#0d854c047001b5465cf1ba21f52904b5" title="cache of iepdf.beta">_beta</a>=elem_mult ( val, ( <a class="code" href="classbdm_1_1migamma.html#c9847093da59a9ba0ebb68d2c592f5dc" title="cache of iepdf.alpha">_alpha</a>-1.0 ) ); 
    595 <a name="l00789"></a>00789                         }; 
    596 <a name="l00790"></a>00790         }; 
    597 <a name="l00791"></a>00791  
    598 <a name="l00792"></a>00792  
    599 <a name="l00804"></a><a class="code" href="classbdm_1_1mgamma__fix.html">00804</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1mgamma__fix.html" title="Gamma random walk around a fixed point.">mgamma_fix</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mgamma.html" title="Gamma random walk.">mgamma</a> 
    600 <a name="l00805"></a>00805         { 
    601 <a name="l00806"></a>00806                 <span class="keyword">protected</span>: 
    602 <a name="l00808"></a><a class="code" href="classbdm_1_1mgamma__fix.html#1eb701506aabb2e6af007e487212d6fa">00808</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1mgamma__fix.html#1eb701506aabb2e6af007e487212d6fa" title="parameter l">l</a>; 
    603 <a name="l00810"></a><a class="code" href="classbdm_1_1mgamma__fix.html#018c6f901a04e419455308a07eb3b0b2">00810</a>                         vec <a class="code" href="classbdm_1_1mgamma__fix.html#018c6f901a04e419455308a07eb3b0b2" title="reference vector">refl</a>; 
    604 <a name="l00811"></a>00811                 <span class="keyword">public</span>: 
    605 <a name="l00813"></a><a class="code" href="classbdm_1_1mgamma__fix.html#9a31bc9b4b60188a18a2a6b588dc4b2d">00813</a>                         <a class="code" href="classbdm_1_1mgamma__fix.html#9a31bc9b4b60188a18a2a6b588dc4b2d" title="Constructor.">mgamma_fix</a> ( ) : <a class="code" href="classbdm_1_1mgamma.html" title="Gamma random walk.">mgamma</a> ( ),<a class="code" href="classbdm_1_1mgamma__fix.html#018c6f901a04e419455308a07eb3b0b2" title="reference vector">refl</a> () {}; 
    606 <a name="l00815"></a><a class="code" href="classbdm_1_1mgamma__fix.html#1bfd30e90db9dc1fbda4a9fbb0b716b2">00815</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgamma__fix.html#1bfd30e90db9dc1fbda4a9fbb0b716b2" title="Set value of k.">set_parameters</a> ( <span class="keywordtype">double</span> k0 , vec ref0, <span class="keywordtype">double</span> l0 ) 
    607 <a name="l00816"></a>00816                         { 
    608 <a name="l00817"></a>00817                                 <a class="code" href="classbdm_1_1mgamma__fix.html#1bfd30e90db9dc1fbda4a9fbb0b716b2" title="Set value of k.">mgamma::set_parameters</a> ( k0, ref0 ); 
    609 <a name="l00818"></a>00818                                 <a class="code" href="classbdm_1_1mgamma__fix.html#018c6f901a04e419455308a07eb3b0b2" title="reference vector">refl</a>=pow ( ref0,1.0-l0 );<a class="code" href="classbdm_1_1mgamma__fix.html#1eb701506aabb2e6af007e487212d6fa" title="parameter l">l</a>=l0; 
    610 <a name="l00819"></a>00819                                 <a class="code" href="classbdm_1_1mpdf.html#7c1900976ff13dbc09c9729b3bbff9e6" title="dimension of the condition">dimc</a>=dimension(); 
    611 <a name="l00820"></a>00820                         }; 
    612 <a name="l00821"></a>00821  
    613 <a name="l00822"></a><a class="code" href="classbdm_1_1mgamma__fix.html#1d539591deb7a38bb3403c2b396c8ff7">00822</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgamma__fix.html#1d539591deb7a38bb3403c2b396c8ff7" title="Update ep so that it represents this mpdf conditioned on rvc = cond.">condition</a> ( <span class="keyword">const</span> vec &amp;val ) {vec mean=elem_mult ( <a class="code" href="classbdm_1_1mgamma__fix.html#018c6f901a04e419455308a07eb3b0b2" title="reference vector">refl</a>,pow ( val,<a class="code" href="classbdm_1_1mgamma__fix.html#1eb701506aabb2e6af007e487212d6fa" title="parameter l">l</a> ) ); <a class="code" href="classbdm_1_1mgamma.html#3d95f4dde9214ff6dba265e18af60312" title="cache of iepdf.beta">_beta</a>=<a class="code" href="classbdm_1_1mgamma.html#b20cf88cca1fe9b0b8f2a412608bfd09" title="Constant .">k</a>/mean;}; 
    614 <a name="l00823"></a>00823         }; 
    615 <a name="l00824"></a>00824  
    616 <a name="l00825"></a>00825  
    617 <a name="l00838"></a><a class="code" href="classbdm_1_1migamma__ref.html">00838</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1migamma__ref.html" title="Inverse-Gamma random walk around a fixed point.">migamma_ref</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1migamma.html" title="Inverse-Gamma random walk.">migamma</a> 
    618 <a name="l00839"></a>00839         { 
    619 <a name="l00840"></a>00840                 <span class="keyword">protected</span>: 
    620 <a name="l00842"></a><a class="code" href="classbdm_1_1migamma__ref.html#cdc1345ba8375fbdb18a69322d2f841d">00842</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1migamma__ref.html#cdc1345ba8375fbdb18a69322d2f841d" title="parameter l">l</a>; 
    621 <a name="l00844"></a><a class="code" href="classbdm_1_1migamma__ref.html#3692dc67caf4367e15564d37f45476f6">00844</a>                         vec <a class="code" href="classbdm_1_1migamma__ref.html#3692dc67caf4367e15564d37f45476f6" title="reference vector">refl</a>; 
    622 <a name="l00845"></a>00845                 <span class="keyword">public</span>: 
    623 <a name="l00847"></a><a class="code" href="classbdm_1_1migamma__ref.html#f45b15a10f084991ba6b48295f10421f">00847</a>                         <a class="code" href="classbdm_1_1migamma__ref.html#f45b15a10f084991ba6b48295f10421f" title="Constructor.">migamma_ref</a> ( ) : <a class="code" href="classbdm_1_1migamma.html" title="Inverse-Gamma random walk.">migamma</a> (),<a class="code" href="classbdm_1_1migamma__ref.html#3692dc67caf4367e15564d37f45476f6" title="reference vector">refl</a> ( ) {}; 
    624 <a name="l00849"></a><a class="code" href="classbdm_1_1migamma__ref.html#b0b4eb278ef5d0831ec4954ba7bd2800">00849</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1migamma__ref.html#b0b4eb278ef5d0831ec4954ba7bd2800" title="Set value of k.">set_parameters</a> ( <span class="keywordtype">double</span> k0 , vec ref0, <span class="keywordtype">double</span> l0 ) 
    625 <a name="l00850"></a>00850                         { 
    626 <a name="l00851"></a>00851                                 <a class="code" href="classbdm_1_1migamma__ref.html#b0b4eb278ef5d0831ec4954ba7bd2800" title="Set value of k.">migamma::set_parameters</a> ( ref0.length(), k0 ); 
    627 <a name="l00852"></a>00852                                 <a class="code" href="classbdm_1_1migamma__ref.html#3692dc67caf4367e15564d37f45476f6" title="reference vector">refl</a>=pow ( ref0,1.0-l0 ); 
    628 <a name="l00853"></a>00853                                 <a class="code" href="classbdm_1_1migamma__ref.html#cdc1345ba8375fbdb18a69322d2f841d" title="parameter l">l</a>=l0; 
    629 <a name="l00854"></a>00854                                 <a class="code" href="classbdm_1_1mpdf.html#7c1900976ff13dbc09c9729b3bbff9e6" title="dimension of the condition">dimc</a> = dimension(); 
    630 <a name="l00855"></a>00855                         }; 
    631 <a name="l00856"></a>00856  
    632 <a name="l00857"></a><a class="code" href="classbdm_1_1migamma__ref.html#ae86b2e4ff963d62e05d4e130514634a">00857</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1migamma__ref.html#ae86b2e4ff963d62e05d4e130514634a" title="Update ep so that it represents this mpdf conditioned on rvc = cond.">condition</a> ( <span class="keyword">const</span> vec &amp;val ) 
    633 <a name="l00858"></a>00858                         { 
    634 <a name="l00859"></a>00859                                 vec mean=elem_mult ( <a class="code" href="classbdm_1_1migamma__ref.html#3692dc67caf4367e15564d37f45476f6" title="reference vector">refl</a>,pow ( val,<a class="code" href="classbdm_1_1migamma__ref.html#cdc1345ba8375fbdb18a69322d2f841d" title="parameter l">l</a> ) ); 
    635 <a name="l00860"></a>00860                                 <a class="code" href="classbdm_1_1migamma__ref.html#ae86b2e4ff963d62e05d4e130514634a" title="Update ep so that it represents this mpdf conditioned on rvc = cond.">migamma::condition</a> ( mean ); 
    636 <a name="l00861"></a>00861                         }; 
     579<a name="l00761"></a><a class="code" href="classbdm_1_1migamma.html">00761</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1migamma.html" title="Inverse-Gamma random walk.">migamma</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mpdf__internal.html" title="Mpdf with internal epdf that is modified by function condition.">mpdf_internal</a>&lt;eigamma&gt; 
     580<a name="l00762"></a>00762 { 
     581<a name="l00763"></a>00763         <span class="keyword">protected</span>: 
     582<a name="l00765"></a><a class="code" href="classbdm_1_1migamma.html#adc56bc9da542e0103ec16b9be8e5e38c">00765</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1migamma.html#adc56bc9da542e0103ec16b9be8e5e38c" title="Constant .">k</a>; 
     583<a name="l00766"></a>00766  
     584<a name="l00768"></a><a class="code" href="classbdm_1_1migamma.html#ac9847093da59a9ba0ebb68d2c592f5dc">00768</a>                 vec &amp;<a class="code" href="classbdm_1_1migamma.html#ac9847093da59a9ba0ebb68d2c592f5dc" title="cache of iepdf.alpha">_alpha</a>; 
     585<a name="l00769"></a>00769  
     586<a name="l00771"></a><a class="code" href="classbdm_1_1migamma.html#a0d854c047001b5465cf1ba21f52904b5">00771</a>                 vec &amp;<a class="code" href="classbdm_1_1migamma.html#a0d854c047001b5465cf1ba21f52904b5" title="cache of iepdf.beta">_beta</a>; 
     587<a name="l00772"></a>00772  
     588<a name="l00773"></a>00773         <span class="keyword">public</span>: 
     589<a name="l00776"></a>00776                 <a class="code" href="classbdm_1_1migamma.html" title="Inverse-Gamma random walk.">migamma</a>() : <a class="code" href="classbdm_1_1mpdf__internal.html" title="Mpdf with internal epdf that is modified by function condition.">mpdf_internal</a>&lt;<a class="code" href="classbdm_1_1eigamma.html" title="Inverse-Gamma posterior density.">eigamma</a>&gt;(), 
     590<a name="l00777"></a>00777                                 <a class="code" href="classbdm_1_1migamma.html#adc56bc9da542e0103ec16b9be8e5e38c" title="Constant .">k</a> (0), 
     591<a name="l00778"></a>00778                                 <a class="code" href="classbdm_1_1migamma.html#ac9847093da59a9ba0ebb68d2c592f5dc" title="cache of iepdf.alpha">_alpha</a> (<a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>.<a class="code" href="classbdm_1_1migamma.html#ac9847093da59a9ba0ebb68d2c592f5dc" title="cache of iepdf.alpha">_alpha</a>()), 
     592<a name="l00779"></a>00779                                 <a class="code" href="classbdm_1_1migamma.html#a0d854c047001b5465cf1ba21f52904b5" title="cache of iepdf.beta">_beta</a> (<a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>.<a class="code" href="classbdm_1_1migamma.html#a0d854c047001b5465cf1ba21f52904b5" title="cache of iepdf.beta">_beta</a>()) { 
     593<a name="l00780"></a>00780                 } 
     594<a name="l00781"></a>00781  
     595<a name="l00782"></a>00782                 <a class="code" href="classbdm_1_1migamma.html" title="Inverse-Gamma random walk.">migamma</a> (<span class="keyword">const</span> <a class="code" href="classbdm_1_1migamma.html" title="Inverse-Gamma random walk.">migamma</a> &amp;m) : <a class="code" href="classbdm_1_1mpdf__internal.html" title="Mpdf with internal epdf that is modified by function condition.">mpdf_internal</a>&lt;<a class="code" href="classbdm_1_1eigamma.html" title="Inverse-Gamma posterior density.">eigamma</a>&gt;(), 
     596<a name="l00783"></a>00783                                 <a class="code" href="classbdm_1_1migamma.html#adc56bc9da542e0103ec16b9be8e5e38c" title="Constant .">k</a> (0), 
     597<a name="l00784"></a>00784                                 <a class="code" href="classbdm_1_1migamma.html#ac9847093da59a9ba0ebb68d2c592f5dc" title="cache of iepdf.alpha">_alpha</a> (<a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>.<a class="code" href="classbdm_1_1migamma.html#ac9847093da59a9ba0ebb68d2c592f5dc" title="cache of iepdf.alpha">_alpha</a>()), 
     598<a name="l00785"></a>00785                                 <a class="code" href="classbdm_1_1migamma.html#a0d854c047001b5465cf1ba21f52904b5" title="cache of iepdf.beta">_beta</a> (<a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>.<a class="code" href="classbdm_1_1migamma.html#a0d854c047001b5465cf1ba21f52904b5" title="cache of iepdf.beta">_beta</a>()) { 
     599<a name="l00786"></a>00786                 } 
     600<a name="l00788"></a>00788  
     601<a name="l00790"></a><a class="code" href="classbdm_1_1migamma.html#a8b10ab922e2a7bae2fb6bb3efc7b6151">00790</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1migamma.html#a8b10ab922e2a7bae2fb6bb3efc7b6151" title="Set value of k.">set_parameters</a> (<span class="keywordtype">int</span> len, <span class="keywordtype">double</span> k0) { 
     602<a name="l00791"></a>00791                         <a class="code" href="classbdm_1_1migamma.html#adc56bc9da542e0103ec16b9be8e5e38c" title="Constant .">k</a> = k0; 
     603<a name="l00792"></a>00792                         <a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>.set_parameters ( (1.0 / (<a class="code" href="classbdm_1_1migamma.html#adc56bc9da542e0103ec16b9be8e5e38c" title="Constant .">k</a>*<a class="code" href="classbdm_1_1migamma.html#adc56bc9da542e0103ec16b9be8e5e38c" title="Constant .">k</a>) + 2.0) *ones (len) <span class="comment">/*alpha*/</span>, ones (len) <span class="comment">/*beta*/</span>); 
     604<a name="l00793"></a>00793                         <a class="code" href="classbdm_1_1mpdf.html#a7c1900976ff13dbc09c9729b3bbff9e6" title="dimension of the condition">dimc</a> = dimension(); 
     605<a name="l00794"></a>00794                 }; 
     606<a name="l00795"></a><a class="code" href="classbdm_1_1migamma.html#a7a34b1e2e3aa2250d7c0ed7df1665b8c">00795</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1migamma.html#a7a34b1e2e3aa2250d7c0ed7df1665b8c">condition</a> (<span class="keyword">const</span> vec &amp;val) { 
     607<a name="l00796"></a>00796                         <a class="code" href="classbdm_1_1migamma.html#a0d854c047001b5465cf1ba21f52904b5" title="cache of iepdf.beta">_beta</a> = elem_mult (val, (<a class="code" href="classbdm_1_1migamma.html#ac9847093da59a9ba0ebb68d2c592f5dc" title="cache of iepdf.alpha">_alpha</a> - 1.0)); 
     608<a name="l00797"></a>00797                 }; 
     609<a name="l00798"></a>00798 }; 
     610<a name="l00799"></a>00799  
     611<a name="l00800"></a>00800  
     612<a name="l00812"></a><a class="code" href="classbdm_1_1mgamma__fix.html">00812</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1mgamma__fix.html" title="Gamma random walk around a fixed point.">mgamma_fix</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mgamma.html" title="Gamma random walk.">mgamma</a> 
     613<a name="l00813"></a>00813 { 
     614<a name="l00814"></a>00814         <span class="keyword">protected</span>: 
     615<a name="l00816"></a><a class="code" href="classbdm_1_1mgamma__fix.html#a1eb701506aabb2e6af007e487212d6fa">00816</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1mgamma__fix.html#a1eb701506aabb2e6af007e487212d6fa" title="parameter l">l</a>; 
     616<a name="l00818"></a><a class="code" href="classbdm_1_1mgamma__fix.html#a018c6f901a04e419455308a07eb3b0b2">00818</a>                 vec <a class="code" href="classbdm_1_1mgamma__fix.html#a018c6f901a04e419455308a07eb3b0b2" title="reference vector">refl</a>; 
     617<a name="l00819"></a>00819         <span class="keyword">public</span>: 
     618<a name="l00821"></a><a class="code" href="classbdm_1_1mgamma__fix.html#a9a31bc9b4b60188a18a2a6b588dc4b2d">00821</a>                 <a class="code" href="classbdm_1_1mgamma__fix.html#a9a31bc9b4b60188a18a2a6b588dc4b2d" title="Constructor.">mgamma_fix</a> () : <a class="code" href="classbdm_1_1mgamma.html" title="Gamma random walk.">mgamma</a> (), <a class="code" href="classbdm_1_1mgamma__fix.html#a018c6f901a04e419455308a07eb3b0b2" title="reference vector">refl</a> () {}; 
     619<a name="l00823"></a><a class="code" href="classbdm_1_1mgamma__fix.html#a1bfd30e90db9dc1fbda4a9fbb0b716b2">00823</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgamma__fix.html#a1bfd30e90db9dc1fbda4a9fbb0b716b2" title="Set value of k.">set_parameters</a> (<span class="keywordtype">double</span> k0 , vec ref0, <span class="keywordtype">double</span> l0) { 
     620<a name="l00824"></a>00824                         <a class="code" href="classbdm_1_1mgamma__fix.html#a1bfd30e90db9dc1fbda4a9fbb0b716b2" title="Set value of k.">mgamma::set_parameters</a> (k0, ref0); 
     621<a name="l00825"></a>00825                         <a class="code" href="classbdm_1_1mgamma__fix.html#a018c6f901a04e419455308a07eb3b0b2" title="reference vector">refl</a> = pow (ref0, 1.0 - l0);<a class="code" href="classbdm_1_1mgamma__fix.html#a1eb701506aabb2e6af007e487212d6fa" title="parameter l">l</a> = l0; 
     622<a name="l00826"></a>00826                         <a class="code" href="classbdm_1_1mpdf.html#a7c1900976ff13dbc09c9729b3bbff9e6" title="dimension of the condition">dimc</a> = dimension(); 
     623<a name="l00827"></a>00827                 }; 
     624<a name="l00828"></a>00828  
     625<a name="l00829"></a><a class="code" href="classbdm_1_1mgamma__fix.html#a1d539591deb7a38bb3403c2b396c8ff7">00829</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgamma__fix.html#a1d539591deb7a38bb3403c2b396c8ff7">condition</a> (<span class="keyword">const</span> vec &amp;val) {vec mean = elem_mult (<a class="code" href="classbdm_1_1mgamma__fix.html#a018c6f901a04e419455308a07eb3b0b2" title="reference vector">refl</a>, pow (val, <a class="code" href="classbdm_1_1mgamma__fix.html#a1eb701506aabb2e6af007e487212d6fa" title="parameter l">l</a>)); <a class="code" href="classbdm_1_1mgamma.html#a3d95f4dde9214ff6dba265e18af60312" title="cache of iepdf.beta">_beta</a> = <a class="code" href="classbdm_1_1mgamma.html#ab20cf88cca1fe9b0b8f2a412608bfd09" title="Constant .">k</a> / mean;}; 
     626<a name="l00830"></a>00830 }; 
     627<a name="l00831"></a>00831  
     628<a name="l00832"></a>00832  
     629<a name="l00845"></a><a class="code" href="classbdm_1_1migamma__ref.html">00845</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1migamma__ref.html" title="Inverse-Gamma random walk around a fixed point.">migamma_ref</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1migamma.html" title="Inverse-Gamma random walk.">migamma</a> 
     630<a name="l00846"></a>00846 { 
     631<a name="l00847"></a>00847         <span class="keyword">protected</span>: 
     632<a name="l00849"></a><a class="code" href="classbdm_1_1migamma__ref.html#acdc1345ba8375fbdb18a69322d2f841d">00849</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1migamma__ref.html#acdc1345ba8375fbdb18a69322d2f841d" title="parameter l">l</a>; 
     633<a name="l00851"></a><a class="code" href="classbdm_1_1migamma__ref.html#a3692dc67caf4367e15564d37f45476f6">00851</a>                 vec <a class="code" href="classbdm_1_1migamma__ref.html#a3692dc67caf4367e15564d37f45476f6" title="reference vector">refl</a>; 
     634<a name="l00852"></a>00852         <span class="keyword">public</span>: 
     635<a name="l00854"></a><a class="code" href="classbdm_1_1migamma__ref.html#af45b15a10f084991ba6b48295f10421f">00854</a>                 <a class="code" href="classbdm_1_1migamma__ref.html#af45b15a10f084991ba6b48295f10421f" title="Constructor.">migamma_ref</a> () : <a class="code" href="classbdm_1_1migamma.html" title="Inverse-Gamma random walk.">migamma</a> (), <a class="code" href="classbdm_1_1migamma__ref.html#a3692dc67caf4367e15564d37f45476f6" title="reference vector">refl</a> () {}; 
     636<a name="l00856"></a><a class="code" href="classbdm_1_1migamma__ref.html#ab0b4eb278ef5d0831ec4954ba7bd2800">00856</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1migamma__ref.html#ab0b4eb278ef5d0831ec4954ba7bd2800" title="Set value of k.">set_parameters</a> (<span class="keywordtype">double</span> k0 , vec ref0, <span class="keywordtype">double</span> l0) { 
     637<a name="l00857"></a>00857                         <a class="code" href="classbdm_1_1migamma__ref.html#ab0b4eb278ef5d0831ec4954ba7bd2800" title="Set value of k.">migamma::set_parameters</a> (ref0.length(), k0); 
     638<a name="l00858"></a>00858                         <a class="code" href="classbdm_1_1migamma__ref.html#a3692dc67caf4367e15564d37f45476f6" title="reference vector">refl</a> = pow (ref0, 1.0 - l0); 
     639<a name="l00859"></a>00859                         <a class="code" href="classbdm_1_1migamma__ref.html#acdc1345ba8375fbdb18a69322d2f841d" title="parameter l">l</a> = l0; 
     640<a name="l00860"></a>00860                         <a class="code" href="classbdm_1_1mpdf.html#a7c1900976ff13dbc09c9729b3bbff9e6" title="dimension of the condition">dimc</a> = dimension(); 
     641<a name="l00861"></a>00861                 }; 
    637642<a name="l00862"></a>00862  
    638 <a name="l00883"></a>00883                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1migamma__ref.html#9e7e0f7d2aa9ecca8ec1af8cbcb5ef1d">from_setting</a>( <span class="keyword">const</span> Setting &amp;<span class="keyword">set</span> ); 
    639 <a name="l00884"></a>00884  
    640 <a name="l00885"></a>00885                         <span class="comment">// TODO dodelat void to_setting( Setting &amp;set ) const;</span> 
    641 <a name="l00886"></a>00886         }; 
    642 <a name="l00887"></a>00887  
    643 <a name="l00888"></a>00888  
    644 <a name="l00889"></a>00889         <a class="code" href="user__info_8h.html#4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a>(migamma_ref); 
    645 <a name="l00890"></a>00890  
    646 <a name="l00900"></a><a class="code" href="classbdm_1_1elognorm.html">00900</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1elognorm.html">elognorm</a>: <span class="keyword">public</span> <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm</a>&lt;ldmat&gt; 
    647 <a name="l00901"></a>00901         { 
    648 <a name="l00902"></a>00902                 <span class="keyword">public</span>: 
    649 <a name="l00903"></a><a class="code" href="classbdm_1_1elognorm.html#8b948e2bce1253765a2542199913aaba">00903</a>                         vec <a class="code" href="classbdm_1_1elognorm.html#8b948e2bce1253765a2542199913aaba" title="Returns a sample,  from density .">sample</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> exp ( <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm&lt;ldmat&gt;::sample</a>() );}; 
    650 <a name="l00904"></a><a class="code" href="classbdm_1_1elognorm.html#adb41e4f4d6600dec6f8c1dbc5ed9eea">00904</a>                         vec <a class="code" href="classbdm_1_1elognorm.html#adb41e4f4d6600dec6f8c1dbc5ed9eea" title="return expected value">mean</a>()<span class="keyword"> const </span>{vec var=<a class="code" href="classbdm_1_1enorm.html#729c75ef0fa8abae03d58ad1f81e6773" title="return expected variance (not covariance!)">enorm&lt;ldmat&gt;::variance</a>();<span class="keywordflow">return</span> exp ( <a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a> - 0.5*var );}; 
    651 <a name="l00905"></a>00905  
    652 <a name="l00906"></a>00906         }; 
    653 <a name="l00907"></a>00907  
    654 <a name="l00919"></a><a class="code" href="classbdm_1_1mlognorm.html">00919</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1mlognorm.html" title="Log-Normal random walk.">mlognorm</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mpdf.html" title="Conditional probability density, e.g. modeling some dependencies.">mpdf</a> 
    655 <a name="l00920"></a>00920         { 
    656 <a name="l00921"></a>00921                 <span class="keyword">protected</span>: 
    657 <a name="l00922"></a>00922                         <a class="code" href="classbdm_1_1shared__ptr.html">shared_ptr&lt;elognorm&gt;</a> eno; 
    658 <a name="l00923"></a>00923  
    659 <a name="l00925"></a><a class="code" href="classbdm_1_1mlognorm.html#a51128a2e503b8b2ce698244b9e0db1a">00925</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1mlognorm.html#a51128a2e503b8b2ce698244b9e0db1a" title="parameter 1/2*sigma^2">sig2</a>; 
    660 <a name="l00926"></a>00926  
    661 <a name="l00928"></a><a class="code" href="classbdm_1_1mlognorm.html#7d0063f77d899ef22e8c5edd642176d2">00928</a>                         vec &amp;<a class="code" href="classbdm_1_1mlognorm.html#7d0063f77d899ef22e8c5edd642176d2" title="access">mu</a>; 
    662 <a name="l00929"></a>00929                 <span class="keyword">public</span>: 
    663 <a name="l00931"></a><a class="code" href="classbdm_1_1mlognorm.html#a5d6eb2688d02e0348b96c4fbd7bde41">00931</a>                         <a class="code" href="classbdm_1_1mlognorm.html#a5d6eb2688d02e0348b96c4fbd7bde41" title="Constructor.">mlognorm</a>():eno(new <a class="code" href="classbdm_1_1elognorm.html">elognorm</a>()), 
    664 <a name="l00932"></a>00932                             <a class="code" href="classbdm_1_1mlognorm.html#a51128a2e503b8b2ce698244b9e0db1a" title="parameter 1/2*sigma^2">sig2</a>(0), 
    665 <a name="l00933"></a>00933                             <a class="code" href="classbdm_1_1mlognorm.html#7d0063f77d899ef22e8c5edd642176d2" title="access">mu</a>(eno-&gt;_mu()) { 
    666 <a name="l00934"></a>00934                             set_ep(eno); 
    667 <a name="l00935"></a>00935                         } 
    668 <a name="l00936"></a>00936  
    669 <a name="l00938"></a><a class="code" href="classbdm_1_1mlognorm.html#604cab0e8a76f9041dc3c606043bb39f">00938</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlognorm.html#604cab0e8a76f9041dc3c606043bb39f" title="Set value of k.">set_parameters</a> ( <span class="keywordtype">int</span> size, <span class="keywordtype">double</span> k ) 
    670 <a name="l00939"></a>00939                         { 
    671 <a name="l00940"></a>00940                                 <a class="code" href="classbdm_1_1mlognorm.html#a51128a2e503b8b2ce698244b9e0db1a" title="parameter 1/2*sigma^2">sig2</a> = 0.5*log ( k*k+1 ); 
    672 <a name="l00941"></a>00941                                 eno-&gt;set_parameters ( zeros ( size ),2*<a class="code" href="classbdm_1_1mlognorm.html#a51128a2e503b8b2ce698244b9e0db1a" title="parameter 1/2*sigma^2">sig2</a>*eye ( size ) ); 
    673 <a name="l00942"></a>00942  
    674 <a name="l00943"></a>00943                                 <a class="code" href="classbdm_1_1mpdf.html#7c1900976ff13dbc09c9729b3bbff9e6" title="dimension of the condition">dimc</a> = size; 
    675 <a name="l00944"></a>00944                         }; 
    676 <a name="l00945"></a>00945  
    677 <a name="l00946"></a><a class="code" href="classbdm_1_1mlognorm.html#9106d8fd8bdf2b6be675ffd8f3ca584e">00946</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlognorm.html#9106d8fd8bdf2b6be675ffd8f3ca584e" title="Update ep so that it represents this mpdf conditioned on rvc = cond.">condition</a> ( <span class="keyword">const</span> vec &amp;val ) 
    678 <a name="l00947"></a>00947                         { 
    679 <a name="l00948"></a>00948                                 <a class="code" href="classbdm_1_1mlognorm.html#7d0063f77d899ef22e8c5edd642176d2" title="access">mu</a>=log ( val )-<a class="code" href="classbdm_1_1mlognorm.html#a51128a2e503b8b2ce698244b9e0db1a" title="parameter 1/2*sigma^2">sig2</a>;<span class="comment">//elem_mult ( refl,pow ( val,l ) );</span> 
    680 <a name="l00949"></a>00949                         }; 
    681 <a name="l00950"></a>00950  
    682 <a name="l00969"></a>00969                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlognorm.html#49e45ea13a869da607ef9be7a229128a">from_setting</a>( <span class="keyword">const</span> Setting &amp;<span class="keyword">set</span> ); 
    683 <a name="l00970"></a>00970  
    684 <a name="l00971"></a>00971                         <span class="comment">// TODO dodelat void to_setting( Setting &amp;set ) const;</span> 
    685 <a name="l00972"></a>00972  
    686 <a name="l00973"></a>00973         }; 
    687 <a name="l00974"></a>00974          
    688 <a name="l00975"></a>00975         <a class="code" href="user__info_8h.html#4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a>(mlognorm); 
    689 <a name="l00976"></a>00976  
    690 <a name="l00980"></a><a class="code" href="classbdm_1_1eWishartCh.html">00980</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1eWishartCh.html">eWishartCh</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> 
    691 <a name="l00981"></a>00981         { 
    692 <a name="l00982"></a>00982                 <span class="keyword">protected</span>: 
    693 <a name="l00984"></a><a class="code" href="classbdm_1_1eWishartCh.html#1b42f9284a32f23b0b253a628cda7490">00984</a>                         <a class="code" href="classchmat.html" title="Symmetric matrix stored in square root decomposition using upper cholesky.">chmat</a> <a class="code" href="classbdm_1_1eWishartCh.html#1b42f9284a32f23b0b253a628cda7490" title="Upper-Triagle of Choleski decomposition of .">Y</a>; 
    694 <a name="l00986"></a><a class="code" href="classbdm_1_1eWishartCh.html#b745c73faef785009484180582050a1f">00986</a>                         <span class="keywordtype">int</span> <a class="code" href="classbdm_1_1eWishartCh.html#b745c73faef785009484180582050a1f" title="dimension of matrix ">p</a>; 
    695 <a name="l00988"></a><a class="code" href="classbdm_1_1eWishartCh.html#1879a14d7d2bb05062523b189baa11c3">00988</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eWishartCh.html#1879a14d7d2bb05062523b189baa11c3" title="degrees of freedom ">delta</a>; 
    696 <a name="l00989"></a>00989                 <span class="keyword">public</span>: 
    697 <a name="l00990"></a>00990                         <span class="keywordtype">void</span> set_parameters ( <span class="keyword">const</span> mat &amp;Y0, <span class="keyword">const</span> <span class="keywordtype">double</span> delta0 ) {<a class="code" href="classbdm_1_1eWishartCh.html#1b42f9284a32f23b0b253a628cda7490" title="Upper-Triagle of Choleski decomposition of .">Y</a>=<a class="code" href="classchmat.html" title="Symmetric matrix stored in square root decomposition using upper cholesky.">chmat</a> ( Y0 );<a class="code" href="classbdm_1_1eWishartCh.html#1879a14d7d2bb05062523b189baa11c3" title="degrees of freedom ">delta</a>=delta0; <a class="code" href="classbdm_1_1eWishartCh.html#b745c73faef785009484180582050a1f" title="dimension of matrix ">p</a>=<a class="code" href="classbdm_1_1eWishartCh.html#1b42f9284a32f23b0b253a628cda7490" title="Upper-Triagle of Choleski decomposition of .">Y</a>.<a class="code" href="classsqmat.html#071e80ced9cc3b8cbb360fa7462eb646" title="Reimplementing common functions of mat: rows().">rows</a>(); <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> = <a class="code" href="classbdm_1_1eWishartCh.html#b745c73faef785009484180582050a1f" title="dimension of matrix ">p</a>*<a class="code" href="classbdm_1_1eWishartCh.html#b745c73faef785009484180582050a1f" title="dimension of matrix ">p</a>; } 
    698 <a name="l00991"></a>00991                         mat sample_mat()<span class="keyword"> const</span> 
    699 <a name="l00992"></a>00992 <span class="keyword">                        </span>{ 
    700 <a name="l00993"></a>00993                                 mat X=zeros ( <a class="code" href="classbdm_1_1eWishartCh.html#b745c73faef785009484180582050a1f" title="dimension of matrix ">p</a>,<a class="code" href="classbdm_1_1eWishartCh.html#b745c73faef785009484180582050a1f" title="dimension of matrix ">p</a> ); 
    701 <a name="l00994"></a>00994  
    702 <a name="l00995"></a>00995                                 <span class="comment">//sample diagonal</span> 
    703 <a name="l00996"></a>00996                                 <span class="keywordflow">for</span> ( <span class="keywordtype">int</span> i=0;i&lt;<a class="code" href="classbdm_1_1eWishartCh.html#b745c73faef785009484180582050a1f" title="dimension of matrix ">p</a>;i++ ) 
    704 <a name="l00997"></a>00997                                 { 
    705 <a name="l00998"></a>00998                                         GamRNG.<a class="code" href="classitpp_1_1Gamma__RNG.html#dfaae19411e39aa87e1f72e409b6babe" title="Set lambda.">setup</a> ( 0.5* ( <a class="code" href="classbdm_1_1eWishartCh.html#1879a14d7d2bb05062523b189baa11c3" title="degrees of freedom ">delta</a>-i ) , 0.5 ); <span class="comment">// no +1 !! index if from 0</span> 
    706 <a name="l00999"></a>00999 <span class="preprocessor">#pragma omp critical</span> 
    707 <a name="l01000"></a>01000 <span class="preprocessor"></span>                                        X ( i,i ) =sqrt ( GamRNG() ); 
    708 <a name="l01001"></a>01001                                 } 
    709 <a name="l01002"></a>01002                                 <span class="comment">//do the rest</span> 
    710 <a name="l01003"></a>01003                                 <span class="keywordflow">for</span> ( <span class="keywordtype">int</span> i=0;i&lt;p;i++ ) 
    711 <a name="l01004"></a>01004                                 { 
    712 <a name="l01005"></a>01005                                         <span class="keywordflow">for</span> ( <span class="keywordtype">int</span> j=i+1;j&lt;p;j++ ) 
    713 <a name="l01006"></a>01006                                         { 
     643<a name="l00863"></a><a class="code" href="classbdm_1_1migamma__ref.html#aae86b2e4ff963d62e05d4e130514634a">00863</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1migamma__ref.html#aae86b2e4ff963d62e05d4e130514634a">condition</a> (<span class="keyword">const</span> vec &amp;val) { 
     644<a name="l00864"></a>00864                         vec mean = elem_mult (<a class="code" href="classbdm_1_1migamma__ref.html#a3692dc67caf4367e15564d37f45476f6" title="reference vector">refl</a>, pow (val, <a class="code" href="classbdm_1_1migamma__ref.html#acdc1345ba8375fbdb18a69322d2f841d" title="parameter l">l</a>)); 
     645<a name="l00865"></a>00865                         <a class="code" href="classbdm_1_1migamma__ref.html#aae86b2e4ff963d62e05d4e130514634a">migamma::condition</a> (mean); 
     646<a name="l00866"></a>00866                 }; 
     647<a name="l00867"></a>00867  
     648<a name="l00888"></a>00888                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1migamma__ref.html#a9e7e0f7d2aa9ecca8ec1af8cbcb5ef1d">from_setting</a> (<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>); 
     649<a name="l00889"></a>00889  
     650<a name="l00890"></a>00890                 <span class="comment">// TODO dodelat void to_setting( Setting &amp;set ) const;</span> 
     651<a name="l00891"></a>00891 }; 
     652<a name="l00892"></a>00892  
     653<a name="l00893"></a>00893  
     654<a name="l00894"></a>00894 <a class="code" href="user__info_8h.html#a4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a> (migamma_ref); 
     655<a name="l00895"></a>00895 SHAREDPTR (migamma_ref); 
     656<a name="l00896"></a>00896  
     657<a name="l00906"></a><a class="code" href="classbdm_1_1elognorm.html">00906</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1elognorm.html">elognorm</a>: <span class="keyword">public</span> <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm</a>&lt;ldmat&gt; 
     658<a name="l00907"></a>00907 { 
     659<a name="l00908"></a>00908         <span class="keyword">public</span>: 
     660<a name="l00909"></a><a class="code" href="classbdm_1_1elognorm.html#a8b948e2bce1253765a2542199913aaba">00909</a>                 vec <a class="code" href="classbdm_1_1elognorm.html#a8b948e2bce1253765a2542199913aaba" title="Returns a sample,  from density .">sample</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> exp (<a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm&lt;ldmat&gt;::sample</a>());}; 
     661<a name="l00910"></a><a class="code" href="classbdm_1_1elognorm.html#aadb41e4f4d6600dec6f8c1dbc5ed9eea">00910</a>                 vec <a class="code" href="classbdm_1_1elognorm.html#aadb41e4f4d6600dec6f8c1dbc5ed9eea" title="return expected value">mean</a>()<span class="keyword"> const </span>{vec var = <a class="code" href="classbdm_1_1enorm.html#a729c75ef0fa8abae03d58ad1f81e6773" title="return expected variance (not covariance!)">enorm&lt;ldmat&gt;::variance</a>();<span class="keywordflow">return</span> exp (<a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a> - 0.5*var);}; 
     662<a name="l00911"></a>00911  
     663<a name="l00912"></a>00912 }; 
     664<a name="l00913"></a>00913  
     665<a name="l00925"></a><a class="code" href="classbdm_1_1mlognorm.html">00925</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1mlognorm.html" title="Log-Normal random walk.">mlognorm</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mpdf__internal.html" title="Mpdf with internal epdf that is modified by function condition.">mpdf_internal</a>&lt;elognorm&gt; 
     666<a name="l00926"></a>00926 { 
     667<a name="l00927"></a>00927         <span class="keyword">protected</span>: 
     668<a name="l00929"></a><a class="code" href="classbdm_1_1mlognorm.html#aa51128a2e503b8b2ce698244b9e0db1a">00929</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1mlognorm.html#aa51128a2e503b8b2ce698244b9e0db1a" title="parameter 1/2*sigma^2">sig2</a>; 
     669<a name="l00930"></a>00930  
     670<a name="l00932"></a><a class="code" href="classbdm_1_1mlognorm.html#a7d0063f77d899ef22e8c5edd642176d2">00932</a>                 vec &amp;<a class="code" href="classbdm_1_1mlognorm.html#a7d0063f77d899ef22e8c5edd642176d2" title="access">mu</a>; 
     671<a name="l00933"></a>00933         <span class="keyword">public</span>: 
     672<a name="l00935"></a><a class="code" href="classbdm_1_1mlognorm.html#aa5d6eb2688d02e0348b96c4fbd7bde41">00935</a>                 <a class="code" href="classbdm_1_1mlognorm.html#aa5d6eb2688d02e0348b96c4fbd7bde41" title="Constructor.">mlognorm</a>() : <a class="code" href="classbdm_1_1mpdf__internal.html" title="Mpdf with internal epdf that is modified by function condition.">mpdf_internal</a>&lt;<a class="code" href="classbdm_1_1elognorm.html">elognorm</a>&gt;(), 
     673<a name="l00936"></a>00936                                 <a class="code" href="classbdm_1_1mlognorm.html#aa51128a2e503b8b2ce698244b9e0db1a" title="parameter 1/2*sigma^2">sig2</a> (0), 
     674<a name="l00937"></a>00937                                 <a class="code" href="classbdm_1_1mlognorm.html#a7d0063f77d899ef22e8c5edd642176d2" title="access">mu</a> (<a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>._mu()) { 
     675<a name="l00938"></a>00938                 } 
     676<a name="l00939"></a>00939  
     677<a name="l00941"></a><a class="code" href="classbdm_1_1mlognorm.html#a604cab0e8a76f9041dc3c606043bb39f">00941</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlognorm.html#a604cab0e8a76f9041dc3c606043bb39f" title="Set value of k.">set_parameters</a> (<span class="keywordtype">int</span> size, <span class="keywordtype">double</span> k) { 
     678<a name="l00942"></a>00942                         <a class="code" href="classbdm_1_1mlognorm.html#aa51128a2e503b8b2ce698244b9e0db1a" title="parameter 1/2*sigma^2">sig2</a> = 0.5 * log (k * k + 1); 
     679<a name="l00943"></a>00943                         <a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>.set_parameters (zeros (size), 2*<a class="code" href="classbdm_1_1mlognorm.html#aa51128a2e503b8b2ce698244b9e0db1a" title="parameter 1/2*sigma^2">sig2</a>*eye (size)); 
     680<a name="l00944"></a>00944  
     681<a name="l00945"></a>00945                         <a class="code" href="classbdm_1_1mpdf.html#a7c1900976ff13dbc09c9729b3bbff9e6" title="dimension of the condition">dimc</a> = size; 
     682<a name="l00946"></a>00946                 }; 
     683<a name="l00947"></a>00947  
     684<a name="l00948"></a><a class="code" href="classbdm_1_1mlognorm.html#a9106d8fd8bdf2b6be675ffd8f3ca584e">00948</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlognorm.html#a9106d8fd8bdf2b6be675ffd8f3ca584e">condition</a> (<span class="keyword">const</span> vec &amp;val) { 
     685<a name="l00949"></a>00949                         <a class="code" href="classbdm_1_1mlognorm.html#a7d0063f77d899ef22e8c5edd642176d2" title="access">mu</a> = log (val) - <a class="code" href="classbdm_1_1mlognorm.html#aa51128a2e503b8b2ce698244b9e0db1a" title="parameter 1/2*sigma^2">sig2</a>;<span class="comment">//elem_mult ( refl,pow ( val,l ) );</span> 
     686<a name="l00950"></a>00950                 }; 
     687<a name="l00951"></a>00951  
     688<a name="l00970"></a>00970                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlognorm.html#a49e45ea13a869da607ef9be7a229128a">from_setting</a> (<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>); 
     689<a name="l00971"></a>00971  
     690<a name="l00972"></a>00972                 <span class="comment">// TODO dodelat void to_setting( Setting &amp;set ) const;</span> 
     691<a name="l00973"></a>00973  
     692<a name="l00974"></a>00974 }; 
     693<a name="l00975"></a>00975  
     694<a name="l00976"></a>00976 <a class="code" href="user__info_8h.html#a4f9de2f17e844047726487b99def99c6" title="Macro for registration of class into map of user-infos, registered class is scriptable...">UIREGISTER</a> (mlognorm); 
     695<a name="l00977"></a>00977 SHAREDPTR (mlognorm); 
     696<a name="l00978"></a>00978  
     697<a name="l00982"></a><a class="code" href="classbdm_1_1eWishartCh.html">00982</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1eWishartCh.html">eWishartCh</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> 
     698<a name="l00983"></a>00983 { 
     699<a name="l00984"></a>00984         <span class="keyword">protected</span>: 
     700<a name="l00986"></a><a class="code" href="classbdm_1_1eWishartCh.html#a1b42f9284a32f23b0b253a628cda7490">00986</a>                 <a class="code" href="classbdm_1_1chmat.html" title="Symmetric matrix stored in square root decomposition using upper cholesky.">chmat</a> <a class="code" href="classbdm_1_1eWishartCh.html#a1b42f9284a32f23b0b253a628cda7490" title="Upper-Triagle of Choleski decomposition of .">Y</a>; 
     701<a name="l00988"></a><a class="code" href="classbdm_1_1eWishartCh.html#ab745c73faef785009484180582050a1f">00988</a>                 <span class="keywordtype">int</span> <a class="code" href="classbdm_1_1eWishartCh.html#ab745c73faef785009484180582050a1f" title="dimension of matrix ">p</a>; 
     702<a name="l00990"></a><a class="code" href="classbdm_1_1eWishartCh.html#a1879a14d7d2bb05062523b189baa11c3">00990</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eWishartCh.html#a1879a14d7d2bb05062523b189baa11c3" title="degrees of freedom ">delta</a>; 
     703<a name="l00991"></a>00991         <span class="keyword">public</span>: 
     704<a name="l00993"></a><a class="code" href="classbdm_1_1eWishartCh.html#a183d961532ee97b7dc5ec81701aa59a0">00993</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eWishartCh.html#a183d961532ee97b7dc5ec81701aa59a0" title="Set internal structures.">set_parameters</a> (<span class="keyword">const</span> mat &amp;Y0, <span class="keyword">const</span> <span class="keywordtype">double</span> delta0) {<a class="code" href="classbdm_1_1eWishartCh.html#a1b42f9284a32f23b0b253a628cda7490" title="Upper-Triagle of Choleski decomposition of .">Y</a> = <a class="code" href="classbdm_1_1chmat.html" title="Symmetric matrix stored in square root decomposition using upper cholesky.">chmat</a> (Y0);<a class="code" href="classbdm_1_1eWishartCh.html#a1879a14d7d2bb05062523b189baa11c3" title="degrees of freedom ">delta</a> = delta0; <a class="code" href="classbdm_1_1eWishartCh.html#ab745c73faef785009484180582050a1f" title="dimension of matrix ">p</a> = <a class="code" href="classbdm_1_1eWishartCh.html#a1b42f9284a32f23b0b253a628cda7490" title="Upper-Triagle of Choleski decomposition of .">Y</a>.<a class="code" href="classbdm_1_1sqmat.html#a73e639221343dcce76c3305524d67590" title="Reimplementing common functions of mat: rows().">rows</a>(); <a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> = <a class="code" href="classbdm_1_1eWishartCh.html#ab745c73faef785009484180582050a1f" title="dimension of matrix ">p</a> * <a class="code" href="classbdm_1_1eWishartCh.html#ab745c73faef785009484180582050a1f" title="dimension of matrix ">p</a>; } 
     705<a name="l00995"></a><a class="code" href="classbdm_1_1eWishartCh.html#acc664035d70d2622bf264a29270f8339">00995</a>                 mat <a class="code" href="classbdm_1_1eWishartCh.html#acc664035d70d2622bf264a29270f8339" title="Sample matrix argument.">sample_mat</a>()<span class="keyword"> const </span>{ 
     706<a name="l00996"></a>00996                         mat X = zeros (<a class="code" href="classbdm_1_1eWishartCh.html#ab745c73faef785009484180582050a1f" title="dimension of matrix ">p</a>, <a class="code" href="classbdm_1_1eWishartCh.html#ab745c73faef785009484180582050a1f" title="dimension of matrix ">p</a>); 
     707<a name="l00997"></a>00997  
     708<a name="l00998"></a>00998                         <span class="comment">//sample diagonal</span> 
     709<a name="l00999"></a>00999                         <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = 0;i &lt; <a class="code" href="classbdm_1_1eWishartCh.html#ab745c73faef785009484180582050a1f" title="dimension of matrix ">p</a>;i++) { 
     710<a name="l01000"></a>01000                                 GamRNG.setup (0.5* (<a class="code" href="classbdm_1_1eWishartCh.html#a1879a14d7d2bb05062523b189baa11c3" title="degrees of freedom ">delta</a> - i) , 0.5);   <span class="comment">// no +1 !! index if from 0</span> 
     711<a name="l01001"></a>01001 <span class="preprocessor">#pragma omp critical</span> 
     712<a name="l01002"></a>01002 <span class="preprocessor"></span>                                X (i, i) = sqrt (GamRNG()); 
     713<a name="l01003"></a>01003                         } 
     714<a name="l01004"></a>01004                         <span class="comment">//do the rest</span> 
     715<a name="l01005"></a>01005                         <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = 0;i &lt; p;i++) { 
     716<a name="l01006"></a>01006                                 <span class="keywordflow">for</span> (<span class="keywordtype">int</span> j = i + 1;j &lt; p;j++) { 
    714717<a name="l01007"></a>01007 <span class="preprocessor">#pragma omp critical</span> 
    715 <a name="l01008"></a>01008 <span class="preprocessor"></span>                                                X ( i,j ) =NorRNG.sample(); 
    716 <a name="l01009"></a>01009                                         } 
    717 <a name="l01010"></a>01010                                 } 
    718 <a name="l01011"></a>01011                                 <span class="keywordflow">return</span> X*<a class="code" href="classbdm_1_1eWishartCh.html#1b42f9284a32f23b0b253a628cda7490" title="Upper-Triagle of Choleski decomposition of .">Y</a>.<a class="code" href="classchmat.html#9c50d31c999d85d8e9d8cf2b69b6ac8c" title="Access function.">_Ch</a>();<span class="comment">// return upper triangular part of the decomposition</span> 
    719 <a name="l01012"></a>01012                         } 
    720 <a name="l01013"></a><a class="code" href="classbdm_1_1eWishartCh.html#8f2154b8b5be8f4c5788f261b6d57b9a">01013</a>                         vec <a class="code" href="classbdm_1_1eWishartCh.html#8f2154b8b5be8f4c5788f261b6d57b9a" title="Returns a sample,  from density .">sample</a> ()<span class="keyword"> const</span> 
    721 <a name="l01014"></a>01014 <span class="keyword">                        </span>{ 
    722 <a name="l01015"></a>01015                                 <span class="keywordflow">return</span> vec ( sample_mat()._data(),p*p ); 
    723 <a name="l01016"></a>01016                         } 
    724 <a name="l01018"></a><a class="code" href="classbdm_1_1eWishartCh.html#4eee757c0535c2a88bb20f0767c64981">01018</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eWishartCh.html#4eee757c0535c2a88bb20f0767c64981" title="fast access function y0 will be copied into Y.Ch.">setY</a> ( <span class="keyword">const</span> mat &amp;Ch0 ) {copy_vector ( <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>,Ch0._data(), <a class="code" href="classbdm_1_1eWishartCh.html#1b42f9284a32f23b0b253a628cda7490" title="Upper-Triagle of Choleski decomposition of .">Y</a>.<a class="code" href="classchmat.html#9c50d31c999d85d8e9d8cf2b69b6ac8c" title="Access function.">_Ch</a>()._data() );} 
    725 <a name="l01020"></a><a class="code" href="classbdm_1_1eWishartCh.html#7eac414ec10b85aa5536b0092c57bc4a">01020</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eWishartCh.html#7eac414ec10b85aa5536b0092c57bc4a" title="fast access function y0 will be copied into Y.Ch.">_setY</a> ( <span class="keyword">const</span> vec &amp;ch0 ) {copy_vector ( <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>, ch0._data(), <a class="code" href="classbdm_1_1eWishartCh.html#1b42f9284a32f23b0b253a628cda7490" title="Upper-Triagle of Choleski decomposition of .">Y</a>.<a class="code" href="classchmat.html#9c50d31c999d85d8e9d8cf2b69b6ac8c" title="Access function.">_Ch</a>()._data() ); } 
    726 <a name="l01022"></a><a class="code" href="classbdm_1_1eWishartCh.html#1708cacb5d8cb1b96395d35f5327cb7e">01022</a>                         <span class="keyword">const</span> <a class="code" href="classchmat.html" title="Symmetric matrix stored in square root decomposition using upper cholesky.">chmat</a>&amp; <a class="code" href="classbdm_1_1eWishartCh.html#1708cacb5d8cb1b96395d35f5327cb7e" title="access function">getY</a>()<span class="keyword">const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1eWishartCh.html#1b42f9284a32f23b0b253a628cda7490" title="Upper-Triagle of Choleski decomposition of .">Y</a>;} 
    727 <a name="l01023"></a>01023         }; 
    728 <a name="l01024"></a>01024  
    729 <a name="l01025"></a>01025         <span class="keyword">class </span>eiWishartCh: <span class="keyword">public</span> epdf 
    730 <a name="l01026"></a>01026         { 
    731 <a name="l01027"></a>01027                 <span class="keyword">protected</span>: 
    732 <a name="l01028"></a>01028                         eWishartCh W; 
    733 <a name="l01029"></a>01029                         <span class="keywordtype">int</span> p; 
    734 <a name="l01030"></a>01030                         <span class="keywordtype">double</span> delta; 
    735 <a name="l01031"></a>01031                 <span class="keyword">public</span>: 
    736 <a name="l01032"></a>01032                         <span class="keywordtype">void</span> set_parameters ( <span class="keyword">const</span> mat &amp;Y0, <span class="keyword">const</span> <span class="keywordtype">double</span> delta0) { 
    737 <a name="l01033"></a>01033                                 delta = delta0; 
    738 <a name="l01034"></a>01034                                 W.set_parameters ( inv ( Y0 ),delta0 );  
    739 <a name="l01035"></a>01035                                 dim = W.dimension(); p=Y0.rows(); 
    740 <a name="l01036"></a>01036                         } 
    741 <a name="l01037"></a>01037                         vec sample()<span class="keyword"> const </span>{mat iCh; iCh=inv ( W.sample_mat() ); <span class="keywordflow">return</span> vec ( iCh._data(),<a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> );} 
    742 <a name="l01038"></a>01038                         <span class="keywordtype">void</span> _setY ( <span class="keyword">const</span> vec &amp;y0 ) 
    743 <a name="l01039"></a>01039                         { 
    744 <a name="l01040"></a>01040                                 mat Ch ( p,p ); 
    745 <a name="l01041"></a>01041                                 mat iCh ( p,p ); 
    746 <a name="l01042"></a>01042                                 copy_vector ( dim, y0._data(), Ch._data() ); 
    747 <a name="l01043"></a>01043                                  
    748 <a name="l01044"></a>01044                                 iCh=inv ( Ch ); 
    749 <a name="l01045"></a>01045                                 W.setY ( iCh ); 
    750 <a name="l01046"></a>01046                         } 
    751 <a name="l01047"></a>01047                         <span class="keyword">virtual</span> <span class="keywordtype">double</span> evallog ( <span class="keyword">const</span> vec &amp;val )<span class="keyword"> const </span>{ 
    752 <a name="l01048"></a>01048                                 <a class="code" href="classchmat.html" title="Symmetric matrix stored in square root decomposition using upper cholesky.">chmat</a> X(p); 
    753 <a name="l01049"></a>01049                                 <span class="keyword">const</span> <a class="code" href="classchmat.html" title="Symmetric matrix stored in square root decomposition using upper cholesky.">chmat</a>&amp; Y=W.getY(); 
    754 <a name="l01050"></a>01050                                   
    755 <a name="l01051"></a>01051                                 copy_vector(p*p,val._data(),X._Ch()._data()); 
    756 <a name="l01052"></a>01052                                 <a class="code" href="classchmat.html" title="Symmetric matrix stored in square root decomposition using upper cholesky.">chmat</a> iX(p);X.inv(iX); 
    757 <a name="l01053"></a>01053                                 <span class="comment">// compute  </span> 
    758 <a name="l01054"></a>01054 <span class="comment">//                              \frac{ |\Psi|^{m/2}|X|^{-(m+p+1)/2}e^{-tr(\Psi X^{-1})/2} }{ 2^{mp/2}\Gamma_p(m/2)},</span> 
    759 <a name="l01055"></a>01055                                 mat M=Y.<a class="code" href="classchmat.html#045addd685f8d978efda232d7dcb070e" title="Conversion to full matrix.">to_mat</a>()*iX.to_mat(); 
    760 <a name="l01056"></a>01056                                  
    761 <a name="l01057"></a>01057                                 <span class="keywordtype">double</span> log1 = 0.5*p*(2*Y.<a class="code" href="classchmat.html#b504ca818203b13e667cb3c503980382" title="Logarithm of a determinant.">logdet</a>())-0.5*(delta+p+1)*(2*X.logdet())-0.5*trace(M);  
    762 <a name="l01058"></a>01058                                 <span class="comment">//Fixme! Multivariate gamma omitted!! it is ok for sampling, but not otherwise!!</span> 
    763 <a name="l01059"></a>01059                                  
    764 <a name="l01060"></a>01060 <span class="comment">/*                              if (0) {</span> 
    765 <a name="l01061"></a>01061 <span class="comment">                                        mat XX=X.to_mat();</span> 
    766 <a name="l01062"></a>01062 <span class="comment">                                        mat YY=Y.to_mat();</span> 
    767 <a name="l01063"></a>01063 <span class="comment">                                        </span> 
    768 <a name="l01064"></a>01064 <span class="comment">                                        double log2 = 0.5*p*log(det(YY))-0.5*(delta+p+1)*log(det(XX))-0.5*trace(YY*inv(XX)); </span> 
    769 <a name="l01065"></a>01065 <span class="comment">                                        cout &lt;&lt; log1 &lt;&lt; "," &lt;&lt; log2 &lt;&lt; endl;</span> 
    770 <a name="l01066"></a>01066 <span class="comment">                                }*/</span> 
    771 <a name="l01067"></a>01067                                 <span class="keywordflow">return</span> log1;                             
    772 <a name="l01068"></a>01068                         }; 
    773 <a name="l01069"></a>01069                          
    774 <a name="l01070"></a>01070         }; 
    775 <a name="l01071"></a>01071  
    776 <a name="l01072"></a>01072         <span class="keyword">class </span>rwiWishartCh : <span class="keyword">public</span> mpdf 
    777 <a name="l01073"></a>01073         { 
    778 <a name="l01074"></a>01074                 <span class="keyword">protected</span>: 
    779 <a name="l01075"></a>01075                         shared_ptr&lt;eiWishartCh&gt; eiW; 
    780 <a name="l01077"></a>01077                         <span class="keywordtype">double</span> sqd; 
    781 <a name="l01078"></a>01078                         <span class="comment">//reference point for diagonal</span> 
    782 <a name="l01079"></a>01079                         vec refl; 
    783 <a name="l01080"></a>01080                         <span class="keywordtype">double</span> l; 
    784 <a name="l01081"></a>01081                         <span class="keywordtype">int</span> p; 
    785 <a name="l01082"></a>01082  
    786 <a name="l01083"></a>01083                 <span class="keyword">public</span>: 
    787 <a name="l01084"></a>01084                         rwiWishartCh():eiW(new eiWishartCh()), 
    788 <a name="l01085"></a>01085                             sqd(0), l(0), p(0) { 
    789 <a name="l01086"></a>01086                             set_ep(eiW); 
    790 <a name="l01087"></a>01087                         } 
    791 <a name="l01088"></a>01088  
    792 <a name="l01089"></a>01089                         <span class="keywordtype">void</span> set_parameters ( <span class="keywordtype">int</span> p0, <span class="keywordtype">double</span> k, vec ref0, <span class="keywordtype">double</span> l0  ) 
    793 <a name="l01090"></a>01090                         { 
    794 <a name="l01091"></a>01091                                 p=p0; 
    795 <a name="l01092"></a>01092                                 <span class="keywordtype">double</span> delta = 2/(k*k)+p+3; 
    796 <a name="l01093"></a>01093                                 sqd=sqrt ( delta-p-1 ); 
    797 <a name="l01094"></a>01094                                 l=l0; 
    798 <a name="l01095"></a>01095                                 refl=pow(ref0,1-l); 
    799 <a name="l01096"></a>01096                                  
    800 <a name="l01097"></a>01097                                 eiW-&gt;set_parameters ( eye ( p ),delta ); 
    801 <a name="l01098"></a>01098                                 <a class="code" href="classbdm_1_1mpdf.html#7c1900976ff13dbc09c9729b3bbff9e6" title="dimension of the condition">dimc</a>=eiW-&gt;dimension(); 
    802 <a name="l01099"></a>01099                         } 
    803 <a name="l01100"></a>01100                         <span class="keywordtype">void</span> condition ( <span class="keyword">const</span> vec &amp;c ) { 
    804 <a name="l01101"></a>01101                                 vec z=c; 
    805 <a name="l01102"></a>01102                                 <span class="keywordtype">int</span> ri=0; 
    806 <a name="l01103"></a>01103                                 <span class="keywordflow">for</span>(<span class="keywordtype">int</span> i=0;i&lt;p*p;i+=(p+1)){<span class="comment">//trace diagonal element</span> 
    807 <a name="l01104"></a>01104                                         z(i) = pow(z(i),l)*refl(ri); 
    808 <a name="l01105"></a>01105                                         ri++; 
    809 <a name="l01106"></a>01106                                 } 
    810 <a name="l01107"></a>01107  
    811 <a name="l01108"></a>01108                                 eiW-&gt;_setY ( sqd*z ); 
    812 <a name="l01109"></a>01109                         } 
    813 <a name="l01110"></a>01110         }; 
     718<a name="l01008"></a>01008 <span class="preprocessor"></span>                                        X (i, j) = NorRNG.sample(); 
     719<a name="l01009"></a>01009                                 } 
     720<a name="l01010"></a>01010                         } 
     721<a name="l01011"></a>01011                         <span class="keywordflow">return</span> X*<a class="code" href="classbdm_1_1eWishartCh.html#a1b42f9284a32f23b0b253a628cda7490" title="Upper-Triagle of Choleski decomposition of .">Y</a>.<a class="code" href="classbdm_1_1chmat.html#a17daa8c5c5914bd3194cb3053c5793a5" title="Access function.">_Ch</a>();<span class="comment">// return upper triangular part of the decomposition</span> 
     722<a name="l01012"></a>01012                 } 
     723<a name="l01013"></a><a class="code" href="classbdm_1_1eWishartCh.html#a8f2154b8b5be8f4c5788f261b6d57b9a">01013</a>                 vec <a class="code" href="classbdm_1_1eWishartCh.html#a8f2154b8b5be8f4c5788f261b6d57b9a" title="Returns a sample,  from density .">sample</a> ()<span class="keyword"> const </span>{ 
     724<a name="l01014"></a>01014                         <span class="keywordflow">return</span> vec (<a class="code" href="classbdm_1_1eWishartCh.html#acc664035d70d2622bf264a29270f8339" title="Sample matrix argument.">sample_mat</a>()._data(), <a class="code" href="classbdm_1_1eWishartCh.html#ab745c73faef785009484180582050a1f" title="dimension of matrix ">p</a>*<a class="code" href="classbdm_1_1eWishartCh.html#ab745c73faef785009484180582050a1f" title="dimension of matrix ">p</a>); 
     725<a name="l01015"></a>01015                 } 
     726<a name="l01017"></a><a class="code" href="classbdm_1_1eWishartCh.html#a4eee757c0535c2a88bb20f0767c64981">01017</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eWishartCh.html#a4eee757c0535c2a88bb20f0767c64981" title="fast access function y0 will be copied into Y.Ch.">setY</a> (<span class="keyword">const</span> mat &amp;Ch0) {copy_vector (<a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>, Ch0._data(), <a class="code" href="classbdm_1_1eWishartCh.html#a1b42f9284a32f23b0b253a628cda7490" title="Upper-Triagle of Choleski decomposition of .">Y</a>.<a class="code" href="classbdm_1_1chmat.html#a17daa8c5c5914bd3194cb3053c5793a5" title="Access function.">_Ch</a>()._data());} 
     727<a name="l01019"></a><a class="code" href="classbdm_1_1eWishartCh.html#a7eac414ec10b85aa5536b0092c57bc4a">01019</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eWishartCh.html#a7eac414ec10b85aa5536b0092c57bc4a" title="fast access function y0 will be copied into Y.Ch.">_setY</a> (<span class="keyword">const</span> vec &amp;ch0) {copy_vector (<a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>, ch0._data(), <a class="code" href="classbdm_1_1eWishartCh.html#a1b42f9284a32f23b0b253a628cda7490" title="Upper-Triagle of Choleski decomposition of .">Y</a>.<a class="code" href="classbdm_1_1chmat.html#a17daa8c5c5914bd3194cb3053c5793a5" title="Access function.">_Ch</a>()._data()); } 
     728<a name="l01021"></a><a class="code" href="classbdm_1_1eWishartCh.html#a1708cacb5d8cb1b96395d35f5327cb7e">01021</a>                 <span class="keyword">const</span> <a class="code" href="classbdm_1_1chmat.html" title="Symmetric matrix stored in square root decomposition using upper cholesky.">chmat</a>&amp; <a class="code" href="classbdm_1_1eWishartCh.html#a1708cacb5d8cb1b96395d35f5327cb7e" title="access function">getY</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1eWishartCh.html#a1b42f9284a32f23b0b253a628cda7490" title="Upper-Triagle of Choleski decomposition of .">Y</a>;} 
     729<a name="l01022"></a>01022 }; 
     730<a name="l01023"></a>01023  
     731<a name="l01025"></a>01025  
     732<a name="l01027"></a><a class="code" href="classbdm_1_1eiWishartCh.html">01027</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1eiWishartCh.html" title="Inverse Wishart on Choleski decomposition.">eiWishartCh</a>: <span class="keyword">public</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> 
     733<a name="l01028"></a>01028 { 
     734<a name="l01029"></a>01029         <span class="keyword">protected</span>: 
     735<a name="l01031"></a><a class="code" href="classbdm_1_1eiWishartCh.html#ac6b684b52dc62b8d46e67d5992f98b9a">01031</a>                 <a class="code" href="classbdm_1_1eWishartCh.html">eWishartCh</a> <a class="code" href="classbdm_1_1eiWishartCh.html#ac6b684b52dc62b8d46e67d5992f98b9a" title="Internal instance of Wishart density.">W</a>; 
     736<a name="l01033"></a><a class="code" href="classbdm_1_1eiWishartCh.html#ac11f1c41183f743b97fbb73a4e9ba6cd">01033</a>                 <span class="keywordtype">int</span> <a class="code" href="classbdm_1_1eiWishartCh.html#ac11f1c41183f743b97fbb73a4e9ba6cd" title="size of Ch">p</a>; 
     737<a name="l01035"></a><a class="code" href="classbdm_1_1eiWishartCh.html#aca3035f7bd5e4597cc1843cf07af8464">01035</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eiWishartCh.html#aca3035f7bd5e4597cc1843cf07af8464" title="parameter delta">delta</a>; 
     738<a name="l01036"></a>01036         <span class="keyword">public</span>: 
     739<a name="l01038"></a><a class="code" href="classbdm_1_1eiWishartCh.html#af463caed9d22d9949f3ee67614e7dcd3">01038</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eiWishartCh.html#af463caed9d22d9949f3ee67614e7dcd3" title="constructor function">set_parameters</a> (<span class="keyword">const</span> mat &amp;Y0, <span class="keyword">const</span> <span class="keywordtype">double</span> delta0) { 
     740<a name="l01039"></a>01039                         <a class="code" href="classbdm_1_1eiWishartCh.html#aca3035f7bd5e4597cc1843cf07af8464" title="parameter delta">delta</a> = delta0; 
     741<a name="l01040"></a>01040                         <a class="code" href="classbdm_1_1eiWishartCh.html#ac6b684b52dc62b8d46e67d5992f98b9a" title="Internal instance of Wishart density.">W</a>.<a class="code" href="classbdm_1_1eWishartCh.html#a183d961532ee97b7dc5ec81701aa59a0" title="Set internal structures.">set_parameters</a> (inv (Y0), delta0); 
     742<a name="l01041"></a>01041                         <a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> = <a class="code" href="classbdm_1_1eiWishartCh.html#ac6b684b52dc62b8d46e67d5992f98b9a" title="Internal instance of Wishart density.">W</a>.<a class="code" href="classbdm_1_1epdf.html#a7083a65f7b7a0d0d13b2c516bd2ec29c" title="Size of the random variable.">dimension</a>(); <a class="code" href="classbdm_1_1eiWishartCh.html#ac11f1c41183f743b97fbb73a4e9ba6cd" title="size of Ch">p</a> = Y0.rows(); 
     743<a name="l01042"></a>01042                 } 
     744<a name="l01043"></a><a class="code" href="classbdm_1_1eiWishartCh.html#a2f668192cc9c2e3a5b7e608164685a3e">01043</a>                 vec <a class="code" href="classbdm_1_1eiWishartCh.html#a2f668192cc9c2e3a5b7e608164685a3e" title="Returns a sample,  from density .">sample</a>()<span class="keyword"> const </span>{mat iCh; iCh = inv (<a class="code" href="classbdm_1_1eiWishartCh.html#ac6b684b52dc62b8d46e67d5992f98b9a" title="Internal instance of Wishart density.">W</a>.<a class="code" href="classbdm_1_1eWishartCh.html#acc664035d70d2622bf264a29270f8339" title="Sample matrix argument.">sample_mat</a>()); <span class="keywordflow">return</span> vec (iCh._data(), <a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>);} 
     745<a name="l01045"></a><a class="code" href="classbdm_1_1eiWishartCh.html#ad9947933dfc94546cbd6a81f95ec5af5">01045</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eiWishartCh.html#ad9947933dfc94546cbd6a81f95ec5af5" title="access function">_setY</a> (<span class="keyword">const</span> vec &amp;y0) { 
     746<a name="l01046"></a>01046                         mat Ch (<a class="code" href="classbdm_1_1eiWishartCh.html#ac11f1c41183f743b97fbb73a4e9ba6cd" title="size of Ch">p</a>, <a class="code" href="classbdm_1_1eiWishartCh.html#ac11f1c41183f743b97fbb73a4e9ba6cd" title="size of Ch">p</a>); 
     747<a name="l01047"></a>01047                         mat iCh (<a class="code" href="classbdm_1_1eiWishartCh.html#ac11f1c41183f743b97fbb73a4e9ba6cd" title="size of Ch">p</a>, <a class="code" href="classbdm_1_1eiWishartCh.html#ac11f1c41183f743b97fbb73a4e9ba6cd" title="size of Ch">p</a>); 
     748<a name="l01048"></a>01048                         copy_vector (<a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>, y0._data(), Ch._data()); 
     749<a name="l01049"></a>01049  
     750<a name="l01050"></a>01050                         iCh = inv (Ch); 
     751<a name="l01051"></a>01051                         <a class="code" href="classbdm_1_1eiWishartCh.html#ac6b684b52dc62b8d46e67d5992f98b9a" title="Internal instance of Wishart density.">W</a>.<a class="code" href="classbdm_1_1eWishartCh.html#a4eee757c0535c2a88bb20f0767c64981" title="fast access function y0 will be copied into Y.Ch.">setY</a> (iCh); 
     752<a name="l01052"></a>01052                 } 
     753<a name="l01053"></a><a class="code" href="classbdm_1_1eiWishartCh.html#aa6ddbd815b8b666dd542e97f009f89bb">01053</a>                 <span class="keyword">virtual</span> <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eiWishartCh.html#aa6ddbd815b8b666dd542e97f009f89bb">evallog</a> (<span class="keyword">const</span> vec &amp;val)<span class="keyword"> const </span>{ 
     754<a name="l01054"></a>01054                         <a class="code" href="classbdm_1_1chmat.html" title="Symmetric matrix stored in square root decomposition using upper cholesky.">chmat</a> X (<a class="code" href="classbdm_1_1eiWishartCh.html#ac11f1c41183f743b97fbb73a4e9ba6cd" title="size of Ch">p</a>); 
     755<a name="l01055"></a>01055                         <span class="keyword">const</span> <a class="code" href="classbdm_1_1chmat.html" title="Symmetric matrix stored in square root decomposition using upper cholesky.">chmat</a>&amp; Y = <a class="code" href="classbdm_1_1eiWishartCh.html#ac6b684b52dc62b8d46e67d5992f98b9a" title="Internal instance of Wishart density.">W</a>.<a class="code" href="classbdm_1_1eWishartCh.html#a1708cacb5d8cb1b96395d35f5327cb7e" title="access function">getY</a>(); 
     756<a name="l01056"></a>01056  
     757<a name="l01057"></a>01057                         copy_vector (<a class="code" href="classbdm_1_1eiWishartCh.html#ac11f1c41183f743b97fbb73a4e9ba6cd" title="size of Ch">p</a>*<a class="code" href="classbdm_1_1eiWishartCh.html#ac11f1c41183f743b97fbb73a4e9ba6cd" title="size of Ch">p</a>, val._data(), X.<a class="code" href="classbdm_1_1chmat.html#a17daa8c5c5914bd3194cb3053c5793a5" title="Access function.">_Ch</a>()._data()); 
     758<a name="l01058"></a>01058                         <a class="code" href="classbdm_1_1chmat.html" title="Symmetric matrix stored in square root decomposition using upper cholesky.">chmat</a> iX (p);X.<a class="code" href="classbdm_1_1chmat.html#acbf3389db96dff41fb2e9532d59b13c0" title="Inversion in the same form, i.e. cholesky.">inv</a> (iX); 
     759<a name="l01059"></a>01059                         <span class="comment">// compute</span> 
     760<a name="l01060"></a>01060 <span class="comment">//                              \frac{ |\Psi|^{m/2}|X|^{-(m+p+1)/2}e^{-tr(\Psi X^{-1})/2} }{ 2^{mp/2}\Gamma_p(m/2)},</span> 
     761<a name="l01061"></a>01061                         mat M = Y.<a class="code" href="classbdm_1_1chmat.html#a4b4c5d4dbb8a3d585b68d936cb6df31b" title="Conversion to full matrix.">to_mat</a>() * iX.to_mat(); 
     762<a name="l01062"></a>01062  
     763<a name="l01063"></a>01063                         <span class="keywordtype">double</span> log1 = 0.5 * p * (2 * Y.<a class="code" href="classbdm_1_1chmat.html#a949ccd174ed19f9cfe36366cbd5c56a4" title="Logarithm of a determinant.">logdet</a>()) - 0.5 * (<a class="code" href="classbdm_1_1eiWishartCh.html#aca3035f7bd5e4597cc1843cf07af8464" title="parameter delta">delta</a> + p + 1) * (2 * X.<a class="code" href="classbdm_1_1chmat.html#a949ccd174ed19f9cfe36366cbd5c56a4" title="Logarithm of a determinant.">logdet</a>()) - 0.5 * trace (M); 
     764<a name="l01064"></a>01064                         <span class="comment">//Fixme! Multivariate gamma omitted!! it is ok for sampling, but not otherwise!!</span> 
     765<a name="l01065"></a>01065  
     766<a name="l01066"></a>01066                         <span class="comment">/*                              if (0) {</span> 
     767<a name="l01067"></a>01067 <span class="comment">                                                                mat XX=X.to_mat();</span> 
     768<a name="l01068"></a>01068 <span class="comment">                                                                mat YY=Y.to_mat();</span> 
     769<a name="l01069"></a>01069 <span class="comment"></span> 
     770<a name="l01070"></a>01070 <span class="comment">                                                                double log2 = 0.5*p*log(det(YY))-0.5*(delta+p+1)*log(det(XX))-0.5*trace(YY*inv(XX));</span> 
     771<a name="l01071"></a>01071 <span class="comment">                                                                cout &lt;&lt; log1 &lt;&lt; &quot;,&quot; &lt;&lt; log2 &lt;&lt; endl;</span> 
     772<a name="l01072"></a>01072 <span class="comment">                                                        }*/</span> 
     773<a name="l01073"></a>01073                         <span class="keywordflow">return</span> log1; 
     774<a name="l01074"></a>01074                 }; 
     775<a name="l01075"></a>01075  
     776<a name="l01076"></a>01076 }; 
     777<a name="l01077"></a>01077  
     778<a name="l01079"></a><a class="code" href="classbdm_1_1rwiWishartCh.html">01079</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1rwiWishartCh.html" title="Random Walk on inverse Wishart.">rwiWishartCh</a> : <span class="keyword">public</span> <a class="code" href="classbdm_1_1mpdf__internal.html" title="Mpdf with internal epdf that is modified by function condition.">mpdf_internal</a>&lt;eiWishartCh&gt; 
     779<a name="l01080"></a>01080 { 
     780<a name="l01081"></a>01081         <span class="keyword">protected</span>: 
     781<a name="l01083"></a><a class="code" href="classbdm_1_1rwiWishartCh.html#a2004675b37f712abe3913057bae97cfb">01083</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1rwiWishartCh.html#a2004675b37f712abe3913057bae97cfb" title="square root of  - needed for computation of  from conditions">sqd</a>; 
     782<a name="l01085"></a><a class="code" href="classbdm_1_1rwiWishartCh.html#a239b26c71ebcc118ba6925dfb6efc50a">01085</a>                 vec <a class="code" href="classbdm_1_1rwiWishartCh.html#a239b26c71ebcc118ba6925dfb6efc50a" title="reference point for diagonal">refl</a>; 
     783<a name="l01087"></a><a class="code" href="classbdm_1_1rwiWishartCh.html#a758cd05efa67c60b4bf4f7554b9ba861">01087</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1rwiWishartCh.html#a758cd05efa67c60b4bf4f7554b9ba861" title="power of the reference">l</a>; 
     784<a name="l01089"></a><a class="code" href="classbdm_1_1rwiWishartCh.html#a7ba06d431995c347d2fce30a59c20663">01089</a>                 <span class="keywordtype">int</span> <a class="code" href="classbdm_1_1rwiWishartCh.html#a7ba06d431995c347d2fce30a59c20663" title="dimension">p</a>; 
     785<a name="l01090"></a>01090  
     786<a name="l01091"></a>01091         <span class="keyword">public</span>: 
     787<a name="l01092"></a>01092                 <a class="code" href="classbdm_1_1rwiWishartCh.html" title="Random Walk on inverse Wishart.">rwiWishartCh</a>() : <a class="code" href="classbdm_1_1rwiWishartCh.html#a2004675b37f712abe3913057bae97cfb" title="square root of  - needed for computation of  from conditions">sqd</a> (0), <a class="code" href="classbdm_1_1rwiWishartCh.html#a758cd05efa67c60b4bf4f7554b9ba861" title="power of the reference">l</a> (0), <a class="code" href="classbdm_1_1rwiWishartCh.html#a7ba06d431995c347d2fce30a59c20663" title="dimension">p</a> (0) {} 
     788<a name="l01094"></a><a class="code" href="classbdm_1_1rwiWishartCh.html#a28549ee3ce1ff8509360171ea3cf717c">01094</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1rwiWishartCh.html#a28549ee3ce1ff8509360171ea3cf717c" title="constructor function">set_parameters</a> (<span class="keywordtype">int</span> p0, <span class="keywordtype">double</span> k, vec ref0, <span class="keywordtype">double</span> l0) { 
     789<a name="l01095"></a>01095                         <a class="code" href="classbdm_1_1rwiWishartCh.html#a7ba06d431995c347d2fce30a59c20663" title="dimension">p</a> = p0; 
     790<a name="l01096"></a>01096                         <span class="keywordtype">double</span> delta = 2 / (k * k) + <a class="code" href="classbdm_1_1rwiWishartCh.html#a7ba06d431995c347d2fce30a59c20663" title="dimension">p</a> + 3; 
     791<a name="l01097"></a>01097                         <a class="code" href="classbdm_1_1rwiWishartCh.html#a2004675b37f712abe3913057bae97cfb" title="square root of  - needed for computation of  from conditions">sqd</a> = sqrt (delta - <a class="code" href="classbdm_1_1rwiWishartCh.html#a7ba06d431995c347d2fce30a59c20663" title="dimension">p</a> - 1); 
     792<a name="l01098"></a>01098                         <a class="code" href="classbdm_1_1rwiWishartCh.html#a758cd05efa67c60b4bf4f7554b9ba861" title="power of the reference">l</a> = l0; 
     793<a name="l01099"></a>01099                         <a class="code" href="classbdm_1_1rwiWishartCh.html#a239b26c71ebcc118ba6925dfb6efc50a" title="reference point for diagonal">refl</a> = pow (ref0, 1 - <a class="code" href="classbdm_1_1rwiWishartCh.html#a758cd05efa67c60b4bf4f7554b9ba861" title="power of the reference">l</a>); 
     794<a name="l01100"></a>01100  
     795<a name="l01101"></a>01101                         <a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>.<a class="code" href="classbdm_1_1eiWishartCh.html#af463caed9d22d9949f3ee67614e7dcd3" title="constructor function">set_parameters</a> (eye (<a class="code" href="classbdm_1_1rwiWishartCh.html#a7ba06d431995c347d2fce30a59c20663" title="dimension">p</a>), delta); 
     796<a name="l01102"></a>01102                         <a class="code" href="classbdm_1_1mpdf.html#a7c1900976ff13dbc09c9729b3bbff9e6" title="dimension of the condition">dimc</a> = <a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>.<a class="code" href="classbdm_1_1epdf.html#a7083a65f7b7a0d0d13b2c516bd2ec29c" title="Size of the random variable.">dimension</a>(); 
     797<a name="l01103"></a>01103                 } 
     798<a name="l01104"></a><a class="code" href="classbdm_1_1rwiWishartCh.html#aac087ba6c885d3faeda9171229f9b4e6">01104</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1rwiWishartCh.html#aac087ba6c885d3faeda9171229f9b4e6">condition</a> (<span class="keyword">const</span> vec &amp;c) { 
     799<a name="l01105"></a>01105                         vec z = c; 
     800<a name="l01106"></a>01106                         <span class="keywordtype">int</span> ri = 0; 
     801<a name="l01107"></a>01107                         <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = 0;i &lt; <a class="code" href="classbdm_1_1rwiWishartCh.html#a7ba06d431995c347d2fce30a59c20663" title="dimension">p</a>*<a class="code" href="classbdm_1_1rwiWishartCh.html#a7ba06d431995c347d2fce30a59c20663" title="dimension">p</a>;i += (p + 1)) {<span class="comment">//trace diagonal element</span> 
     802<a name="l01108"></a>01108                                 z (i) = pow (z (i), <a class="code" href="classbdm_1_1rwiWishartCh.html#a758cd05efa67c60b4bf4f7554b9ba861" title="power of the reference">l</a>) * <a class="code" href="classbdm_1_1rwiWishartCh.html#a239b26c71ebcc118ba6925dfb6efc50a" title="reference point for diagonal">refl</a> (ri); 
     803<a name="l01109"></a>01109                                 ri++; 
     804<a name="l01110"></a>01110                         } 
    814805<a name="l01111"></a>01111  
    815 <a name="l01113"></a>01113         <span class="keyword">enum</span> RESAMPLING_METHOD { MULTINOMIAL = 0, STRATIFIED = 1, SYSTEMATIC = 3 }; 
    816 <a name="l01119"></a><a class="code" href="classbdm_1_1eEmp.html">01119</a>         <span class="keyword">class </span><a class="code" href="classbdm_1_1eEmp.html" title="Weighted empirical density.">eEmp</a>: <span class="keyword">public</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> 
    817 <a name="l01120"></a>01120         { 
    818 <a name="l01121"></a>01121                 <span class="keyword">protected</span> : 
    819 <a name="l01123"></a><a class="code" href="classbdm_1_1eEmp.html#9798006271ca77629855113f1283a031">01123</a>                         <span class="keywordtype">int</span> <a class="code" href="classbdm_1_1eEmp.html#9798006271ca77629855113f1283a031" title="Number of particles.">n</a>; 
    820 <a name="l01125"></a><a class="code" href="classbdm_1_1eEmp.html#9d39241aab7c4bbeb07c6d516421c67d">01125</a>                         vec <a class="code" href="classbdm_1_1eEmp.html#9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a>; 
    821 <a name="l01127"></a><a class="code" href="classbdm_1_1eEmp.html#73d819553a0f268b055a087d2d4486f3">01127</a>                         Array&lt;vec&gt; <a class="code" href="classbdm_1_1eEmp.html#73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a>; 
    822 <a name="l01128"></a>01128                 <span class="keyword">public</span>: 
    823 <a name="l01131"></a>01131                         <a class="code" href="classbdm_1_1eEmp.html" title="Weighted empirical density.">eEmp</a> ( ) :<a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> ( ),<a class="code" href="classbdm_1_1eEmp.html#9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a> ( ),<a class="code" href="classbdm_1_1eEmp.html#73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> ( ) {}; 
    824 <a name="l01133"></a><a class="code" href="classbdm_1_1eEmp.html#a3daf6363455af099921715e1233c076">01133</a>                         <a class="code" href="classbdm_1_1eEmp.html#a3daf6363455af099921715e1233c076" title="copy constructor">eEmp</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1eEmp.html" title="Weighted empirical density.">eEmp</a> &amp;e ) : <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> ( e ), <a class="code" href="classbdm_1_1eEmp.html#9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a> ( e.<a class="code" href="classbdm_1_1eEmp.html#9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a> ), <a class="code" href="classbdm_1_1eEmp.html#73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> ( e.<a class="code" href="classbdm_1_1eEmp.html#73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> ) {}; 
    825 <a name="l01135"></a>01135  
    826 <a name="l01137"></a>01137                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eEmp.html#7cfd383180b486fe4526bdf0179350c0" title="Set samples and weights.">set_statistics</a> ( <span class="keyword">const</span> vec &amp;w0, <span class="keyword">const</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a>* pdf0 ); 
    827 <a name="l01139"></a><a class="code" href="classbdm_1_1eEmp.html#cef74aa5f87d10d440b9b1e8bc78c1e7">01139</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eEmp.html#cef74aa5f87d10d440b9b1e8bc78c1e7" title="Set samples and weights.">set_statistics</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a>* pdf0 , <span class="keywordtype">int</span> <a class="code" href="classbdm_1_1eEmp.html#9798006271ca77629855113f1283a031" title="Number of particles.">n</a> ) {<a class="code" href="classbdm_1_1eEmp.html#cef74aa5f87d10d440b9b1e8bc78c1e7" title="Set samples and weights.">set_statistics</a> ( ones ( n ) /n,pdf0 );}; 
    828 <a name="l01141"></a>01141                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eEmp.html#b62d802b8ef39f7c4dcbeb366c90951a" title="Set sample.">set_samples</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a>* pdf0 ); 
    829 <a name="l01143"></a><a class="code" href="classbdm_1_1eEmp.html#c74c281d652356c19b6b079e42ca7ef1">01143</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eEmp.html#c74c281d652356c19b6b079e42ca7ef1" title="Set sample.">set_parameters</a> ( <span class="keywordtype">int</span> n0, <span class="keywordtype">bool</span> copy=<span class="keyword">true</span> ) {<a class="code" href="classbdm_1_1eEmp.html#9798006271ca77629855113f1283a031" title="Number of particles.">n</a>=n0; <a class="code" href="classbdm_1_1eEmp.html#9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a>.set_size ( n0,copy );<a class="code" href="classbdm_1_1eEmp.html#73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a>.set_size ( n0,copy );}; 
    830 <a name="l01145"></a><a class="code" href="classbdm_1_1eEmp.html#d7f83cc0415cd44ae7cc8b4bdad93aef">01145</a>                         vec&amp; <a class="code" href="classbdm_1_1eEmp.html#d7f83cc0415cd44ae7cc8b4bdad93aef" title="Potentially dangerous, use with care.">_w</a>()  {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1eEmp.html#9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a>;}; 
    831 <a name="l01147"></a><a class="code" href="classbdm_1_1eEmp.html#b7d7106f486e3fad38590914a693d714">01147</a>                         <span class="keyword">const</span> vec&amp; <a class="code" href="classbdm_1_1eEmp.html#b7d7106f486e3fad38590914a693d714" title="Potentially dangerous, use with care.">_w</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1eEmp.html#9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a>;}; 
    832 <a name="l01149"></a><a class="code" href="classbdm_1_1eEmp.html#c24966b0aaeb767bc8a6b4fd60931be2">01149</a>                         Array&lt;vec&gt;&amp; <a class="code" href="classbdm_1_1eEmp.html#c24966b0aaeb767bc8a6b4fd60931be2" title="access function">_samples</a>() {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1eEmp.html#73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a>;}; 
    833 <a name="l01151"></a><a class="code" href="classbdm_1_1eEmp.html#b59af0efdb009d98ea8ebfa965e74ae2">01151</a>                         <span class="keyword">const</span> Array&lt;vec&gt;&amp; <a class="code" href="classbdm_1_1eEmp.html#b59af0efdb009d98ea8ebfa965e74ae2" title="access function">_samples</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1eEmp.html#73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a>;}; 
    834 <a name="l01153"></a>01153                         ivec <a class="code" href="classbdm_1_1eEmp.html#f06ce255de5dbb2313f52ee51f82ba3d" title="Function performs resampling, i.e. removal of low-weight samples and duplication...">resample</a> ( RESAMPLING_METHOD method=SYSTEMATIC ); 
    835 <a name="l01155"></a><a class="code" href="classbdm_1_1eEmp.html#97f1e07b5ae6eebc91c7365f0f88d270">01155</a>                         vec <a class="code" href="classbdm_1_1eEmp.html#97f1e07b5ae6eebc91c7365f0f88d270" title="inherited operation : NOT implemneted">sample</a>()<span class="keyword"> const </span>{it_error ( <span class="stringliteral">"Not implemented"</span> );<span class="keywordflow">return</span> 0;} 
    836 <a name="l01157"></a><a class="code" href="classbdm_1_1eEmp.html#01654c014d3aa068f8d4ecba4be86d09">01157</a>                         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eEmp.html#01654c014d3aa068f8d4ecba4be86d09" title="inherited operation : NOT implemneted">evallog</a> ( <span class="keyword">const</span> vec &amp;val )<span class="keyword"> const </span>{it_error ( <span class="stringliteral">"Not implemented"</span> );<span class="keywordflow">return</span> 0.0;} 
    837 <a name="l01158"></a><a class="code" href="classbdm_1_1eEmp.html#bbfcb4f868c7381298c281a256d8c4b9">01158</a>                         vec <a class="code" href="classbdm_1_1eEmp.html#bbfcb4f868c7381298c281a256d8c4b9" title="return expected value">mean</a>()<span class="keyword"> const</span> 
    838 <a name="l01159"></a>01159 <span class="keyword">                        </span>{ 
    839 <a name="l01160"></a>01160                                 vec pom=zeros ( <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> ); 
    840 <a name="l01161"></a>01161                                 <span class="keywordflow">for</span> ( <span class="keywordtype">int</span> i=0;i&lt;<a class="code" href="classbdm_1_1eEmp.html#9798006271ca77629855113f1283a031" title="Number of particles.">n</a>;i++ ) {pom+=<a class="code" href="classbdm_1_1eEmp.html#73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> ( i ) *<a class="code" href="classbdm_1_1eEmp.html#9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a> ( i );} 
    841 <a name="l01162"></a>01162                                 <span class="keywordflow">return</span> pom; 
    842 <a name="l01163"></a>01163                         } 
    843 <a name="l01164"></a><a class="code" href="classbdm_1_1eEmp.html#05e9ebf467ede737cb6a3621d7fd3c87">01164</a>                         vec <a class="code" href="classbdm_1_1eEmp.html#05e9ebf467ede737cb6a3621d7fd3c87" title="return expected variance (not covariance!)">variance</a>()<span class="keyword"> const</span> 
    844 <a name="l01165"></a>01165 <span class="keyword">                        </span>{ 
    845 <a name="l01166"></a>01166                                 vec pom=zeros ( <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> ); 
    846 <a name="l01167"></a>01167                                 <span class="keywordflow">for</span> ( <span class="keywordtype">int</span> i=0;i&lt;<a class="code" href="classbdm_1_1eEmp.html#9798006271ca77629855113f1283a031" title="Number of particles.">n</a>;i++ ) {pom+=pow ( <a class="code" href="classbdm_1_1eEmp.html#73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> ( i ),2 ) *<a class="code" href="classbdm_1_1eEmp.html#9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a> ( i );} 
    847 <a name="l01168"></a>01168                                 <span class="keywordflow">return</span> pom-pow ( <a class="code" href="classbdm_1_1eEmp.html#bbfcb4f868c7381298c281a256d8c4b9" title="return expected value">mean</a>(),2 ); 
    848 <a name="l01169"></a>01169                         } 
    849 <a name="l01171"></a><a class="code" href="classbdm_1_1eEmp.html#b1c9df656144edf79ba2d885613f661f">01171</a>                         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eEmp.html#b1c9df656144edf79ba2d885613f661f" title="For this class, qbounds are minimum and maximum value of the population!">qbounds</a> ( vec &amp;lb, vec &amp;ub, <span class="keywordtype">double</span> perc=0.95 )<span class="keyword"> const</span> 
    850 <a name="l01172"></a>01172 <span class="keyword">                        </span>{ 
    851 <a name="l01173"></a>01173                                 <span class="comment">// lb in inf so than it will be pushed below;</span> 
    852 <a name="l01174"></a>01174                                 lb.set_size ( <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> ); 
    853 <a name="l01175"></a>01175                                 ub.set_size ( <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> ); 
    854 <a name="l01176"></a>01176                                 lb = std::numeric_limits&lt;double&gt;::infinity(); 
    855 <a name="l01177"></a>01177                                 ub = -std::numeric_limits&lt;double&gt;::infinity(); 
    856 <a name="l01178"></a>01178                                 <span class="keywordtype">int</span> j; 
    857 <a name="l01179"></a>01179                                 <span class="keywordflow">for</span> ( <span class="keywordtype">int</span> i=0;i&lt;<a class="code" href="classbdm_1_1eEmp.html#9798006271ca77629855113f1283a031" title="Number of particles.">n</a>;i++ ) 
    858 <a name="l01180"></a>01180                                 { 
    859 <a name="l01181"></a>01181                                         <span class="keywordflow">for</span> ( j=0;j&lt;<a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>; j++ ) 
    860 <a name="l01182"></a>01182                                         { 
    861 <a name="l01183"></a>01183                                                 <span class="keywordflow">if</span> ( <a class="code" href="classbdm_1_1eEmp.html#73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> ( i ) ( j ) &lt;lb ( j ) ) {lb ( j ) =<a class="code" href="classbdm_1_1eEmp.html#73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> ( i ) ( j );} 
    862 <a name="l01184"></a>01184                                                 <span class="keywordflow">if</span> ( <a class="code" href="classbdm_1_1eEmp.html#73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> ( i ) ( j ) &gt;ub ( j ) ) {ub ( j ) =<a class="code" href="classbdm_1_1eEmp.html#73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> ( i ) ( j );} 
    863 <a name="l01185"></a>01185                                         } 
    864 <a name="l01186"></a>01186                                 } 
    865 <a name="l01187"></a>01187                         } 
    866 <a name="l01188"></a>01188         }; 
    867 <a name="l01189"></a>01189  
    868 <a name="l01190"></a>01190  
    869 <a name="l01192"></a>01192  
    870 <a name="l01193"></a>01193         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    871 <a name="l01194"></a>01194         <span class="keywordtype">void</span> enorm&lt;sq_T&gt;::set_parameters ( <span class="keyword">const</span> vec &amp;mu0, <span class="keyword">const</span> sq_T &amp;R0 ) 
    872 <a name="l01195"></a>01195         { 
    873 <a name="l01196"></a>01196 <span class="comment">//Fixme test dimensions of mu0 and R0;</span> 
    874 <a name="l01197"></a>01197                 mu = mu0; 
    875 <a name="l01198"></a>01198                 R = R0; 
    876 <a name="l01199"></a>01199                 <a class="code" href="classbdm_1_1root.html#1c314bd6d6dacb8ba78ea5eb88fd9516" title="This method TODO.">validate</a>(); 
    877 <a name="l01200"></a>01200         }; 
    878 <a name="l01201"></a>01201  
    879 <a name="l01202"></a>01202         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    880 <a name="l01203"></a><a class="code" href="classbdm_1_1enorm.html#61bd470764020bea6e1ed35000f259e6">01203</a>         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1enorm.html#61bd470764020bea6e1ed35000f259e6">enorm&lt;sq_T&gt;::from_setting</a>(<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>){ 
    881 <a name="l01204"></a>01204                 <a class="code" href="classbdm_1_1enorm.html#61bd470764020bea6e1ed35000f259e6">epdf::from_setting</a>(<span class="keyword">set</span>); <span class="comment">//reads rv</span> 
    882 <a name="l01205"></a>01205                  
    883 <a name="l01206"></a>01206                 <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>(<a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a>,<span class="keyword">set</span>,<span class="stringliteral">"mu"</span>, UI::compulsory); 
    884 <a name="l01207"></a>01207                 mat Rtmp;<span class="comment">// necessary for conversion</span> 
    885 <a name="l01208"></a>01208                 <a class="code" href="classbdm_1_1UI.html#cd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a>(Rtmp,<span class="keyword">set</span>,<span class="stringliteral">"R"</span>, UI::compulsory); 
    886 <a name="l01209"></a>01209                 <a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>=Rtmp; <span class="comment">// conversion</span> 
    887 <a name="l01210"></a>01210                 <a class="code" href="classbdm_1_1enorm.html#38eb17ecb75d94a50b6782fcf735cfea" title="This method TODO.">validate</a>(); 
    888 <a name="l01211"></a>01211         } 
    889 <a name="l01212"></a>01212  
    890 <a name="l01213"></a>01213         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    891 <a name="l01214"></a><a class="code" href="classbdm_1_1enorm.html#d2e0d3a1e30ab3ab04df2d0c43ae74a2">01214</a>         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1enorm.html#d2e0d3a1e30ab3ab04df2d0c43ae74a2" title="dupdate in exponential form (not really handy)">enorm&lt;sq_T&gt;::dupdate</a> ( mat &amp;v, <span class="keywordtype">double</span> nu ) 
    892 <a name="l01215"></a>01215         { 
    893 <a name="l01216"></a>01216                 <span class="comment">//</span> 
    894 <a name="l01217"></a>01217         }; 
    895 <a name="l01218"></a>01218  
    896 <a name="l01219"></a>01219 <span class="comment">// template&lt;class sq_T&gt;</span> 
    897 <a name="l01220"></a>01220 <span class="comment">// void enorm&lt;sq_T&gt;::tupdate ( double phi, mat &amp;vbar, double nubar ) {</span> 
    898 <a name="l01221"></a>01221 <span class="comment">//      //</span> 
    899 <a name="l01222"></a>01222 <span class="comment">// };</span> 
    900 <a name="l01223"></a>01223  
    901 <a name="l01224"></a>01224         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    902 <a name="l01225"></a><a class="code" href="classbdm_1_1enorm.html#e1a48f52351ec3a349bd443b713b1766">01225</a>         vec <a class="code" href="classbdm_1_1enorm.html#e1a48f52351ec3a349bd443b713b1766" title="Returns a sample,  from density .">enorm&lt;sq_T&gt;::sample</a>()<span class="keyword"> const</span> 
    903 <a name="l01226"></a>01226 <span class="keyword">        </span>{ 
    904 <a name="l01227"></a>01227                 vec x ( <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a> ); 
    905 <a name="l01228"></a>01228 <span class="preprocessor">#pragma omp critical</span> 
    906 <a name="l01229"></a>01229 <span class="preprocessor"></span>                NorRNG.sample_vector ( <a class="code" href="classbdm_1_1epdf.html#16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>,x ); 
    907 <a name="l01230"></a>01230                 vec smp = <a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>.sqrt_mult ( x ); 
    908 <a name="l01231"></a>01231  
    909 <a name="l01232"></a>01232                 smp += <a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a>; 
    910 <a name="l01233"></a>01233                 <span class="keywordflow">return</span> smp; 
    911 <a name="l01234"></a>01234         }; 
     806<a name="l01112"></a>01112                         <a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>.<a class="code" href="classbdm_1_1eiWishartCh.html#ad9947933dfc94546cbd6a81f95ec5af5" title="access function">_setY</a> (<a class="code" href="classbdm_1_1rwiWishartCh.html#a2004675b37f712abe3913057bae97cfb" title="square root of  - needed for computation of  from conditions">sqd</a>*z); 
     807<a name="l01113"></a>01113                 } 
     808<a name="l01114"></a>01114 }; 
     809<a name="l01115"></a>01115  
     810<a name="l01117"></a>01117 <span class="keyword">enum</span> RESAMPLING_METHOD { MULTINOMIAL = 0, STRATIFIED = 1, SYSTEMATIC = 3 }; 
     811<a name="l01123"></a><a class="code" href="classbdm_1_1eEmp.html">01123</a> <span class="keyword">class </span><a class="code" href="classbdm_1_1eEmp.html" title="Weighted empirical density.">eEmp</a>: <span class="keyword">public</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> 
     812<a name="l01124"></a>01124 { 
     813<a name="l01125"></a>01125         <span class="keyword">protected</span> : 
     814<a name="l01127"></a><a class="code" href="classbdm_1_1eEmp.html#a9798006271ca77629855113f1283a031">01127</a>                 <span class="keywordtype">int</span> <a class="code" href="classbdm_1_1eEmp.html#a9798006271ca77629855113f1283a031" title="Number of particles.">n</a>; 
     815<a name="l01129"></a><a class="code" href="classbdm_1_1eEmp.html#a9d39241aab7c4bbeb07c6d516421c67d">01129</a>                 vec <a class="code" href="classbdm_1_1eEmp.html#a9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a>; 
     816<a name="l01131"></a><a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3">01131</a>                 Array&lt;vec&gt; <a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a>; 
     817<a name="l01132"></a>01132         <span class="keyword">public</span>: 
     818<a name="l01135"></a>01135                 <a class="code" href="classbdm_1_1eEmp.html" title="Weighted empirical density.">eEmp</a> () : <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> (), <a class="code" href="classbdm_1_1eEmp.html#a9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a> (), <a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> () {}; 
     819<a name="l01137"></a><a class="code" href="classbdm_1_1eEmp.html#aa3daf6363455af099921715e1233c076">01137</a>                 <a class="code" href="classbdm_1_1eEmp.html#aa3daf6363455af099921715e1233c076" title="copy constructor">eEmp</a> (<span class="keyword">const</span> <a class="code" href="classbdm_1_1eEmp.html" title="Weighted empirical density.">eEmp</a> &amp;e) : <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> (e), <a class="code" href="classbdm_1_1eEmp.html#a9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a> (e.<a class="code" href="classbdm_1_1eEmp.html#a9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a>), <a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> (e.<a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a>) {}; 
     820<a name="l01139"></a>01139  
     821<a name="l01141"></a>01141                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eEmp.html#ae7f8f98310c1de51bd5c8a1c87528f72" title="Set samples and weights.">set_statistics</a> (<span class="keyword">const</span> vec &amp;w0, <span class="keyword">const</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> &amp;pdf0); 
     822<a name="l01143"></a><a class="code" href="classbdm_1_1eEmp.html#a4f7e6ba7183972e3c7ae399613861b95">01143</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eEmp.html#a4f7e6ba7183972e3c7ae399613861b95" title="Set samples and weights.">set_statistics</a> (<span class="keyword">const</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a> &amp;pdf0 , <span class="keywordtype">int</span> <a class="code" href="classbdm_1_1eEmp.html#a9798006271ca77629855113f1283a031" title="Number of particles.">n</a>) {<a class="code" href="classbdm_1_1eEmp.html#a4f7e6ba7183972e3c7ae399613861b95" title="Set samples and weights.">set_statistics</a> (ones (n) / n, pdf0);}; 
     823<a name="l01145"></a>01145                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eEmp.html#ab62d802b8ef39f7c4dcbeb366c90951a" title="Set sample.">set_samples</a> (<span class="keyword">const</span> <a class="code" href="classbdm_1_1epdf.html" title="Probability density function with numerical statistics, e.g. posterior density.">epdf</a>* pdf0); 
     824<a name="l01147"></a><a class="code" href="classbdm_1_1eEmp.html#ac74c281d652356c19b6b079e42ca7ef1">01147</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eEmp.html#ac74c281d652356c19b6b079e42ca7ef1" title="Set sample.">set_parameters</a> (<span class="keywordtype">int</span> n0, <span class="keywordtype">bool</span> copy = <span class="keyword">true</span>) {<a class="code" href="classbdm_1_1eEmp.html#a9798006271ca77629855113f1283a031" title="Number of particles.">n</a> = n0; <a class="code" href="classbdm_1_1eEmp.html#a9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a>.set_size (n0, copy);<a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a>.set_size (n0, copy);}; 
     825<a name="l01149"></a><a class="code" href="classbdm_1_1eEmp.html#a37937007a3a676cf653a7de412674481">01149</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eEmp.html#a37937007a3a676cf653a7de412674481" title="Set samples.">set_parameters</a> (<span class="keyword">const</span> Array&lt;vec&gt; &amp;Av) { 
     826<a name="l01150"></a>01150                         <a class="code" href="bdmerror_8h.html#a89a0f906b242b79c5d3d342291b2cab4" title="Throw std::runtime_exception if t is not true and NDEBUG is not defined.">bdm_assert_debug</a>(Av.size()&gt;0,<span class="stringliteral">&quot;Empty samples&quot;</span>);  
     827<a name="l01151"></a>01151                         <a class="code" href="classbdm_1_1eEmp.html#a9798006271ca77629855113f1283a031" title="Number of particles.">n</a> = Av.size();  
     828<a name="l01152"></a>01152                         <a class="code" href="classbdm_1_1eEmp.html#ac74c281d652356c19b6b079e42ca7ef1" title="Set sample.">epdf::set_parameters</a>(Av(0).length()); 
     829<a name="l01153"></a>01153                         <a class="code" href="classbdm_1_1eEmp.html#a9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a>=1/n*ones(n); 
     830<a name="l01154"></a>01154                         <a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a>=Av; 
     831<a name="l01155"></a>01155                 }; 
     832<a name="l01157"></a><a class="code" href="classbdm_1_1eEmp.html#ad7f83cc0415cd44ae7cc8b4bdad93aef">01157</a>                 vec&amp; <a class="code" href="classbdm_1_1eEmp.html#ad7f83cc0415cd44ae7cc8b4bdad93aef" title="Potentially dangerous, use with care.">_w</a>()  {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1eEmp.html#a9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a>;}; 
     833<a name="l01159"></a><a class="code" href="classbdm_1_1eEmp.html#ab7d7106f486e3fad38590914a693d714">01159</a>                 <span class="keyword">const</span> vec&amp; <a class="code" href="classbdm_1_1eEmp.html#ab7d7106f486e3fad38590914a693d714" title="Potentially dangerous, use with care.">_w</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1eEmp.html#a9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a>;}; 
     834<a name="l01161"></a><a class="code" href="classbdm_1_1eEmp.html#ac24966b0aaeb767bc8a6b4fd60931be2">01161</a>                 Array&lt;vec&gt;&amp; <a class="code" href="classbdm_1_1eEmp.html#ac24966b0aaeb767bc8a6b4fd60931be2" title="access function">_samples</a>() {<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a>;}; 
     835<a name="l01163"></a><a class="code" href="classbdm_1_1eEmp.html#ab59af0efdb009d98ea8ebfa965e74ae2">01163</a>                 <span class="keyword">const</span> Array&lt;vec&gt;&amp; <a class="code" href="classbdm_1_1eEmp.html#ab59af0efdb009d98ea8ebfa965e74ae2" title="access function">_samples</a>()<span class="keyword"> const </span>{<span class="keywordflow">return</span> <a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a>;}; 
     836<a name="l01165"></a>01165                 ivec <a class="code" href="classbdm_1_1eEmp.html#af06ce255de5dbb2313f52ee51f82ba3d" title="Function performs resampling, i.e. removal of low-weight samples and duplication...">resample</a> (RESAMPLING_METHOD method = SYSTEMATIC); 
     837<a name="l01166"></a>01166  
     838<a name="l01168"></a><a class="code" href="classbdm_1_1eEmp.html#a97f1e07b5ae6eebc91c7365f0f88d270">01168</a>                 vec <a class="code" href="classbdm_1_1eEmp.html#a97f1e07b5ae6eebc91c7365f0f88d270" title="inherited operation : NOT implemented">sample</a>()<span class="keyword"> const </span>{ 
     839<a name="l01169"></a>01169                         <a class="code" href="bdmerror_8h.html#a7c43f3a72afe68ab0c85663a1bb3521a" title="Unconditionally throw std::runtime_error.">bdm_error</a> (<span class="stringliteral">&quot;Not implemented&quot;</span>); 
     840<a name="l01170"></a>01170                         <span class="keywordflow">return</span> vec(); 
     841<a name="l01171"></a>01171                 } 
     842<a name="l01172"></a>01172  
     843<a name="l01174"></a><a class="code" href="classbdm_1_1eEmp.html#a01654c014d3aa068f8d4ecba4be86d09">01174</a>                 <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1eEmp.html#a01654c014d3aa068f8d4ecba4be86d09" title="inherited operation : NOT implemented">evallog</a> (<span class="keyword">const</span> vec &amp;val)<span class="keyword"> const </span>{ 
     844<a name="l01175"></a>01175                         <a class="code" href="bdmerror_8h.html#a7c43f3a72afe68ab0c85663a1bb3521a" title="Unconditionally throw std::runtime_error.">bdm_error</a> (<span class="stringliteral">&quot;Not implemented&quot;</span>); 
     845<a name="l01176"></a>01176                         <span class="keywordflow">return</span> 0.0; 
     846<a name="l01177"></a>01177                 } 
     847<a name="l01178"></a>01178  
     848<a name="l01179"></a><a class="code" href="classbdm_1_1eEmp.html#abbfcb4f868c7381298c281a256d8c4b9">01179</a>                 vec <a class="code" href="classbdm_1_1eEmp.html#abbfcb4f868c7381298c281a256d8c4b9" title="return expected value">mean</a>()<span class="keyword"> const </span>{ 
     849<a name="l01180"></a>01180                         vec pom = zeros (<a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>); 
     850<a name="l01181"></a>01181                         <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = 0;i &lt; <a class="code" href="classbdm_1_1eEmp.html#a9798006271ca77629855113f1283a031" title="Number of particles.">n</a>;i++) {pom += <a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> (i) * <a class="code" href="classbdm_1_1eEmp.html#a9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a> (i);} 
     851<a name="l01182"></a>01182                         <span class="keywordflow">return</span> pom; 
     852<a name="l01183"></a>01183                 } 
     853<a name="l01184"></a><a class="code" href="classbdm_1_1eEmp.html#a05e9ebf467ede737cb6a3621d7fd3c87">01184</a>                 vec <a class="code" href="classbdm_1_1eEmp.html#a05e9ebf467ede737cb6a3621d7fd3c87" title="return expected variance (not covariance!)">variance</a>()<span class="keyword"> const </span>{ 
     854<a name="l01185"></a>01185                         vec pom = zeros (<a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>); 
     855<a name="l01186"></a>01186                         <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = 0;i &lt; <a class="code" href="classbdm_1_1eEmp.html#a9798006271ca77629855113f1283a031" title="Number of particles.">n</a>;i++) {pom += pow (<a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> (i), 2) * <a class="code" href="classbdm_1_1eEmp.html#a9d39241aab7c4bbeb07c6d516421c67d" title="Sample weights .">w</a> (i);} 
     856<a name="l01187"></a>01187                         <span class="keywordflow">return</span> pom -pow (<a class="code" href="classbdm_1_1eEmp.html#abbfcb4f868c7381298c281a256d8c4b9" title="return expected value">mean</a>(), 2); 
     857<a name="l01188"></a>01188                 } 
     858<a name="l01190"></a><a class="code" href="classbdm_1_1eEmp.html#ab1c9df656144edf79ba2d885613f661f">01190</a>                 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1eEmp.html#ab1c9df656144edf79ba2d885613f661f" title="For this class, qbounds are minimum and maximum value of the population!">qbounds</a> (vec &amp;lb, vec &amp;ub, <span class="keywordtype">double</span> perc = 0.95)<span class="keyword"> const </span>{ 
     859<a name="l01191"></a>01191                         <span class="comment">// lb in inf so than it will be pushed below;</span> 
     860<a name="l01192"></a>01192                         lb.set_size (<a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>); 
     861<a name="l01193"></a>01193                         ub.set_size (<a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>); 
     862<a name="l01194"></a>01194                         lb = std::numeric_limits&lt;double&gt;::infinity(); 
     863<a name="l01195"></a>01195                         ub = -std::numeric_limits&lt;double&gt;::infinity(); 
     864<a name="l01196"></a>01196                         <span class="keywordtype">int</span> j; 
     865<a name="l01197"></a>01197                         <span class="keywordflow">for</span> (<span class="keywordtype">int</span> i = 0;i &lt; <a class="code" href="classbdm_1_1eEmp.html#a9798006271ca77629855113f1283a031" title="Number of particles.">n</a>;i++) { 
     866<a name="l01198"></a>01198                                 <span class="keywordflow">for</span> (j = 0;j &lt; <a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>; j++) { 
     867<a name="l01199"></a>01199                                         <span class="keywordflow">if</span> (<a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> (i) (j) &lt; lb (j)) {lb (j) = <a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> (i) (j);} 
     868<a name="l01200"></a>01200                                         <span class="keywordflow">if</span> (<a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> (i) (j) &gt; ub (j)) {ub (j) = <a class="code" href="classbdm_1_1eEmp.html#a73d819553a0f268b055a087d2d4486f3" title="Samples .">samples</a> (i) (j);} 
     869<a name="l01201"></a>01201                                 } 
     870<a name="l01202"></a>01202                         } 
     871<a name="l01203"></a>01203                 } 
     872<a name="l01204"></a>01204 }; 
     873<a name="l01205"></a>01205  
     874<a name="l01206"></a>01206  
     875<a name="l01208"></a>01208  
     876<a name="l01209"></a>01209 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     877<a name="l01210"></a>01210 <span class="keywordtype">void</span> enorm&lt;sq_T&gt;::set_parameters (<span class="keyword">const</span> vec &amp;mu0, <span class="keyword">const</span> sq_T &amp;R0) 
     878<a name="l01211"></a>01211 { 
     879<a name="l01212"></a>01212 <span class="comment">//Fixme test dimensions of mu0 and R0;</span> 
     880<a name="l01213"></a>01213         mu = mu0; 
     881<a name="l01214"></a>01214         R = R0; 
     882<a name="l01215"></a>01215         <a class="code" href="classbdm_1_1root.html#a1c314bd6d6dacb8ba78ea5eb88fd9516" title="This method TODO.">validate</a>(); 
     883<a name="l01216"></a>01216 }; 
     884<a name="l01217"></a>01217  
     885<a name="l01218"></a>01218 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     886<a name="l01219"></a><a class="code" href="classbdm_1_1enorm.html#a61bd470764020bea6e1ed35000f259e6">01219</a> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1enorm.html#a61bd470764020bea6e1ed35000f259e6">enorm&lt;sq_T&gt;::from_setting</a> (<span class="keyword">const</span> Setting &amp;<span class="keyword">set</span>) 
     887<a name="l01220"></a>01220 { 
     888<a name="l01221"></a>01221         <a class="code" href="classbdm_1_1enorm.html#a61bd470764020bea6e1ed35000f259e6">epdf::from_setting</a> (<span class="keyword">set</span>); <span class="comment">//reads rv</span> 
     889<a name="l01222"></a>01222  
     890<a name="l01223"></a>01223         <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (<a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a>, <span class="keyword">set</span>, <span class="stringliteral">&quot;mu&quot;</span>, UI::compulsory); 
     891<a name="l01224"></a>01224         mat Rtmp;<span class="comment">// necessary for conversion</span> 
     892<a name="l01225"></a>01225         <a class="code" href="classbdm_1_1UI.html#acd1667e6fec99ec64dabcb3ca2ff922d">UI::get</a> (Rtmp, <span class="keyword">set</span>, <span class="stringliteral">&quot;R&quot;</span>, UI::compulsory); 
     893<a name="l01226"></a>01226         <a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a> = Rtmp; <span class="comment">// conversion</span> 
     894<a name="l01227"></a>01227         <a class="code" href="classbdm_1_1enorm.html#a38eb17ecb75d94a50b6782fcf735cfea" title="This method TODO.">validate</a>(); 
     895<a name="l01228"></a>01228 } 
     896<a name="l01229"></a>01229  
     897<a name="l01230"></a>01230 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     898<a name="l01231"></a><a class="code" href="classbdm_1_1enorm.html#ad2e0d3a1e30ab3ab04df2d0c43ae74a2">01231</a> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1enorm.html#ad2e0d3a1e30ab3ab04df2d0c43ae74a2" title="dupdate in exponential form (not really handy)">enorm&lt;sq_T&gt;::dupdate</a> (mat &amp;v, <span class="keywordtype">double</span> nu) 
     899<a name="l01232"></a>01232 { 
     900<a name="l01233"></a>01233         <span class="comment">//</span> 
     901<a name="l01234"></a>01234 }; 
    912902<a name="l01235"></a>01235  
    913903<a name="l01236"></a>01236 <span class="comment">// template&lt;class sq_T&gt;</span> 
    914 <a name="l01237"></a>01237 <span class="comment">// double enorm&lt;sq_T&gt;::eval ( const vec &amp;val ) const {</span> 
    915 <a name="l01238"></a>01238 <span class="comment">//      double pdfl,e;</span> 
    916 <a name="l01239"></a>01239 <span class="comment">//      pdfl = evallog ( val );</span> 
    917 <a name="l01240"></a>01240 <span class="comment">//      e = exp ( pdfl );</span> 
    918 <a name="l01241"></a>01241 <span class="comment">//      return e;</span> 
    919 <a name="l01242"></a>01242 <span class="comment">// };</span> 
    920 <a name="l01243"></a>01243  
    921 <a name="l01244"></a>01244         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    922 <a name="l01245"></a><a class="code" href="classbdm_1_1enorm.html#e13aeed5b543b2179bacdc4fa2ae47a3">01245</a>         <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1enorm.html#e13aeed5b543b2179bacdc4fa2ae47a3" title="Evaluate normalized log-probability.">enorm&lt;sq_T&gt;::evallog_nn</a> ( <span class="keyword">const</span> vec &amp;val )<span class="keyword"> const</span> 
    923 <a name="l01246"></a>01246 <span class="keyword">        </span>{ 
    924 <a name="l01247"></a>01247                 <span class="comment">// 1.83787706640935 = log(2pi)</span> 
    925 <a name="l01248"></a>01248                 <span class="keywordtype">double</span> tmp=-0.5* ( <a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>.invqform ( <a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a>-val ) );<span class="comment">// - lognc();</span> 
    926 <a name="l01249"></a>01249                 <span class="keywordflow">return</span>  tmp; 
    927 <a name="l01250"></a>01250         }; 
    928 <a name="l01251"></a>01251  
    929 <a name="l01252"></a>01252         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    930 <a name="l01253"></a><a class="code" href="classbdm_1_1enorm.html#25785343aff102cc5df1cab08ba16d32">01253</a>         <span class="keyword">inline</span> <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1enorm.html#25785343aff102cc5df1cab08ba16d32" title="logarithm of the normalizing constant, ">enorm&lt;sq_T&gt;::lognc</a> ()<span class="keyword"> const</span> 
    931 <a name="l01254"></a>01254 <span class="keyword">        </span>{ 
    932 <a name="l01255"></a>01255                 <span class="comment">// 1.83787706640935 = log(2pi)</span> 
    933 <a name="l01256"></a>01256                 <span class="keywordtype">double</span> tmp=0.5* ( <a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>.cols() * 1.83787706640935 +<a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>.logdet() ); 
    934 <a name="l01257"></a>01257                 <span class="keywordflow">return</span> tmp; 
    935 <a name="l01258"></a>01258         }; 
    936 <a name="l01259"></a>01259  
    937 <a name="l01260"></a>01260         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    938 <a name="l01261"></a><a class="code" href="classbdm_1_1mlnorm.html#5d18dec3167584338a4775c1d165d96f">01261</a>         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlnorm.html#5d18dec3167584338a4775c1d165d96f" title="Set A and R.">mlnorm&lt;sq_T&gt;::set_parameters</a> ( <span class="keyword">const</span> mat &amp;A0, <span class="keyword">const</span> vec &amp;mu0, <span class="keyword">const</span> sq_T &amp;R0 ) 
    939 <a name="l01262"></a>01262         { 
    940 <a name="l01263"></a>01263                 it_assert_debug ( A0.rows() ==mu0.length(),<span class="stringliteral">""</span> ); 
    941 <a name="l01264"></a>01264                 it_assert_debug ( A0.rows() ==R0.rows(),<span class="stringliteral">""</span> ); 
    942 <a name="l01265"></a>01265  
    943 <a name="l01266"></a>01266                 <a class="code" href="classbdm_1_1mlnorm.html#6de45ab731143e70a94318df586c36ba" title="Internal epdf that arise by conditioning on rvc.">iepdf</a>-&gt;set_parameters(zeros(A0.rows()), R0); 
    944 <a name="l01267"></a>01267                 A = A0; 
    945 <a name="l01268"></a>01268                 mu_const = mu0; 
    946 <a name="l01269"></a>01269                 <a class="code" href="classbdm_1_1mpdf.html#7c1900976ff13dbc09c9729b3bbff9e6" title="dimension of the condition">dimc</a> = A0.cols(); 
    947 <a name="l01270"></a>01270         } 
    948 <a name="l01271"></a>01271  
    949 <a name="l01272"></a>01272 <span class="comment">// template&lt;class sq_T&gt;</span> 
    950 <a name="l01273"></a>01273 <span class="comment">// vec mlnorm&lt;sq_T&gt;::samplecond (const  vec &amp;cond, double &amp;lik ) {</span> 
    951 <a name="l01274"></a>01274 <span class="comment">//      this-&gt;condition ( cond );</span> 
    952 <a name="l01275"></a>01275 <span class="comment">//      vec smp = epdf.sample();</span> 
    953 <a name="l01276"></a>01276 <span class="comment">//      lik = epdf.eval ( smp );</span> 
    954 <a name="l01277"></a>01277 <span class="comment">//      return smp;</span> 
    955 <a name="l01278"></a>01278 <span class="comment">// }</span> 
    956 <a name="l01279"></a>01279  
    957 <a name="l01280"></a>01280 <span class="comment">// template&lt;class sq_T&gt;</span> 
    958 <a name="l01281"></a>01281 <span class="comment">// mat mlnorm&lt;sq_T&gt;::samplecond (const vec &amp;cond, vec &amp;lik, int n ) {</span> 
    959 <a name="l01282"></a>01282 <span class="comment">//      int i;</span> 
    960 <a name="l01283"></a>01283 <span class="comment">//      int dim = rv.count();</span> 
    961 <a name="l01284"></a>01284 <span class="comment">//      mat Smp ( dim,n );</span> 
    962 <a name="l01285"></a>01285 <span class="comment">//      vec smp ( dim );</span> 
    963 <a name="l01286"></a>01286 <span class="comment">//      this-&gt;condition ( cond );</span> 
    964 <a name="l01287"></a>01287 <span class="comment">//</span> 
    965 <a name="l01288"></a>01288 <span class="comment">//      for ( i=0; i&lt;n; i++ ) {</span> 
    966 <a name="l01289"></a>01289 <span class="comment">//              smp = epdf.sample();</span> 
    967 <a name="l01290"></a>01290 <span class="comment">//              lik ( i ) = epdf.eval ( smp );</span> 
    968 <a name="l01291"></a>01291 <span class="comment">//              Smp.set_col ( i ,smp );</span> 
    969 <a name="l01292"></a>01292 <span class="comment">//      }</span> 
     904<a name="l01237"></a>01237 <span class="comment">// void enorm&lt;sq_T&gt;::tupdate ( double phi, mat &amp;vbar, double nubar ) {</span> 
     905<a name="l01238"></a>01238 <span class="comment">//      //</span> 
     906<a name="l01239"></a>01239 <span class="comment">// };</span> 
     907<a name="l01240"></a>01240  
     908<a name="l01241"></a>01241 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     909<a name="l01242"></a><a class="code" href="classbdm_1_1enorm.html#ae1a48f52351ec3a349bd443b713b1766">01242</a> vec <a class="code" href="classbdm_1_1enorm.html#ae1a48f52351ec3a349bd443b713b1766" title="Returns a sample,  from density .">enorm&lt;sq_T&gt;::sample</a>()<span class="keyword"> const</span> 
     910<a name="l01243"></a>01243 <span class="keyword"></span>{ 
     911<a name="l01244"></a>01244         vec x (<a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>); 
     912<a name="l01245"></a>01245 <span class="preprocessor">#pragma omp critical</span> 
     913<a name="l01246"></a>01246 <span class="preprocessor"></span>        NorRNG.sample_vector (<a class="code" href="classbdm_1_1epdf.html#a16adac20ec7fe07e1ea0b27d917788ce" title="dimension of the random variable">dim</a>, x); 
     914<a name="l01247"></a>01247         vec smp = <a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>.sqrt_mult (x); 
     915<a name="l01248"></a>01248  
     916<a name="l01249"></a>01249         smp += <a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a>; 
     917<a name="l01250"></a>01250         <span class="keywordflow">return</span> smp; 
     918<a name="l01251"></a>01251 }; 
     919<a name="l01252"></a>01252  
     920<a name="l01253"></a>01253 <span class="comment">// template&lt;class sq_T&gt;</span> 
     921<a name="l01254"></a>01254 <span class="comment">// double enorm&lt;sq_T&gt;::eval ( const vec &amp;val ) const {</span> 
     922<a name="l01255"></a>01255 <span class="comment">//      double pdfl,e;</span> 
     923<a name="l01256"></a>01256 <span class="comment">//      pdfl = evallog ( val );</span> 
     924<a name="l01257"></a>01257 <span class="comment">//      e = exp ( pdfl );</span> 
     925<a name="l01258"></a>01258 <span class="comment">//      return e;</span> 
     926<a name="l01259"></a>01259 <span class="comment">// };</span> 
     927<a name="l01260"></a>01260  
     928<a name="l01261"></a>01261 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     929<a name="l01262"></a><a class="code" href="classbdm_1_1enorm.html#ae13aeed5b543b2179bacdc4fa2ae47a3">01262</a> <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1enorm.html#ae13aeed5b543b2179bacdc4fa2ae47a3" title="Evaluate normalized log-probability.">enorm&lt;sq_T&gt;::evallog_nn</a> (<span class="keyword">const</span> vec &amp;val)<span class="keyword"> const</span> 
     930<a name="l01263"></a>01263 <span class="keyword"></span>{ 
     931<a name="l01264"></a>01264         <span class="comment">// 1.83787706640935 = log(2pi)</span> 
     932<a name="l01265"></a>01265         <span class="keywordtype">double</span> tmp = -0.5 * (<a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>.invqform (<a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a> - val));<span class="comment">// - lognc();</span> 
     933<a name="l01266"></a>01266         <span class="keywordflow">return</span>  tmp; 
     934<a name="l01267"></a>01267 }; 
     935<a name="l01268"></a>01268  
     936<a name="l01269"></a>01269 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     937<a name="l01270"></a><a class="code" href="classbdm_1_1enorm.html#a25785343aff102cc5df1cab08ba16d32">01270</a> <span class="keyword">inline</span> <span class="keywordtype">double</span> <a class="code" href="classbdm_1_1enorm.html#a25785343aff102cc5df1cab08ba16d32" title="logarithm of the normalizing constant, ">enorm&lt;sq_T&gt;::lognc</a> ()<span class="keyword"> const</span> 
     938<a name="l01271"></a>01271 <span class="keyword"></span>{ 
     939<a name="l01272"></a>01272         <span class="comment">// 1.83787706640935 = log(2pi)</span> 
     940<a name="l01273"></a>01273         <span class="keywordtype">double</span> tmp = 0.5 * (<a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>.cols() * 1.83787706640935 + <a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>.logdet()); 
     941<a name="l01274"></a>01274         <span class="keywordflow">return</span> tmp; 
     942<a name="l01275"></a>01275 }; 
     943<a name="l01276"></a>01276  
     944<a name="l01277"></a>01277  
     945<a name="l01278"></a>01278 <span class="comment">// template&lt;class sq_T&gt;</span> 
     946<a name="l01279"></a>01279 <span class="comment">// vec mlnorm&lt;sq_T&gt;::samplecond (const  vec &amp;cond, double &amp;lik ) {</span> 
     947<a name="l01280"></a>01280 <span class="comment">//      this-&gt;condition ( cond );</span> 
     948<a name="l01281"></a>01281 <span class="comment">//      vec smp = epdf.sample();</span> 
     949<a name="l01282"></a>01282 <span class="comment">//      lik = epdf.eval ( smp );</span> 
     950<a name="l01283"></a>01283 <span class="comment">//      return smp;</span> 
     951<a name="l01284"></a>01284 <span class="comment">// }</span> 
     952<a name="l01285"></a>01285  
     953<a name="l01286"></a>01286 <span class="comment">// template&lt;class sq_T&gt;</span> 
     954<a name="l01287"></a>01287 <span class="comment">// mat mlnorm&lt;sq_T&gt;::samplecond (const vec &amp;cond, vec &amp;lik, int n ) {</span> 
     955<a name="l01288"></a>01288 <span class="comment">//      int i;</span> 
     956<a name="l01289"></a>01289 <span class="comment">//      int dim = rv.count();</span> 
     957<a name="l01290"></a>01290 <span class="comment">//      mat Smp ( dim,n );</span> 
     958<a name="l01291"></a>01291 <span class="comment">//      vec smp ( dim );</span> 
     959<a name="l01292"></a>01292 <span class="comment">//      this-&gt;condition ( cond );</span> 
    970960<a name="l01293"></a>01293 <span class="comment">//</span> 
    971 <a name="l01294"></a>01294 <span class="comment">//      return Smp;</span> 
    972 <a name="l01295"></a>01295 <span class="comment">// }</span> 
    973 <a name="l01296"></a>01296  
    974 <a name="l01297"></a>01297         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    975 <a name="l01298"></a><a class="code" href="classbdm_1_1mlnorm.html#0dafc0196e7e07fd06dc6716e0e18bc2">01298</a>         <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mlnorm.html#0dafc0196e7e07fd06dc6716e0e18bc2">mlnorm&lt;sq_T&gt;::condition</a> ( <span class="keyword">const</span> vec &amp;cond ) 
    976 <a name="l01299"></a>01299         { 
    977 <a name="l01300"></a>01300                 _mu = A*cond + mu_const; 
    978 <a name="l01301"></a>01301 <span class="comment">//R is already assigned;</span> 
    979 <a name="l01302"></a>01302         } 
     961<a name="l01294"></a>01294 <span class="comment">//      for ( i=0; i&lt;n; i++ ) {</span> 
     962<a name="l01295"></a>01295 <span class="comment">//              smp = epdf.sample();</span> 
     963<a name="l01296"></a>01296 <span class="comment">//              lik ( i ) = epdf.eval ( smp );</span> 
     964<a name="l01297"></a>01297 <span class="comment">//              Smp.set_col ( i ,smp );</span> 
     965<a name="l01298"></a>01298 <span class="comment">//      }</span> 
     966<a name="l01299"></a>01299 <span class="comment">//</span> 
     967<a name="l01300"></a>01300 <span class="comment">//      return Smp;</span> 
     968<a name="l01301"></a>01301 <span class="comment">// }</span> 
     969<a name="l01302"></a>01302  
    980970<a name="l01303"></a>01303  
    981 <a name="l01304"></a>01304         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    982 <a name="l01305"></a><a class="code" href="classbdm_1_1enorm.html#c2996bdaffad38fbe0fc776db54c9d80">01305</a>         <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm&lt;sq_T&gt;</a>* <a class="code" href="classbdm_1_1enorm.html#c2996bdaffad38fbe0fc776db54c9d80" title="Return marginal density on the given RV, the remainig rvs are intergrated out.">enorm&lt;sq_T&gt;::marginal</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a> &amp;rvn )<span class="keyword"> const</span> 
    983 <a name="l01306"></a>01306 <span class="keyword">        </span>{ 
    984 <a name="l01307"></a>01307                 it_assert_debug ( <a class="code" href="classbdm_1_1epdf.html#c4b863ff84c7a4882fb3ad18556027f9" title="True if rv is assigned.">isnamed</a>(), <span class="stringliteral">"rv description is not assigned"</span> ); 
    985 <a name="l01308"></a>01308                 ivec irvn = rvn.<a class="code" href="classbdm_1_1RV.html#cbebdb5e0d30101a6eb63550ef701c55">dataind</a> ( <a class="code" href="classbdm_1_1epdf.html#62c5b8ff71d9ebe6cd58d3c342eb1dc8" title="Description of the random variable.">rv</a> ); 
    986 <a name="l01309"></a>01309  
    987 <a name="l01310"></a>01310                 sq_T Rn ( <a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>,irvn ); <span class="comment">//select rows and columns of R</span> 
    988 <a name="l01311"></a>01311  
    989 <a name="l01312"></a>01312                 <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm&lt;sq_T&gt;</a>* tmp = <span class="keyword">new</span> <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm&lt;sq_T&gt;</a>; 
    990 <a name="l01313"></a>01313                 tmp-&gt;<a class="code" href="classbdm_1_1epdf.html#f423e28448dbb69ef4905295ec8de8ff" title="Name its rv.">set_rv</a> ( rvn ); 
    991 <a name="l01314"></a>01314                 tmp-&gt;set_parameters ( <a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a> ( irvn ), Rn ); 
    992 <a name="l01315"></a>01315                 <span class="keywordflow">return</span> tmp; 
    993 <a name="l01316"></a>01316         } 
    994 <a name="l01317"></a>01317  
    995 <a name="l01318"></a>01318         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    996 <a name="l01319"></a><a class="code" href="classbdm_1_1enorm.html#baea4d49c657342b58297d68cda16d26">01319</a>         <a class="code" href="classbdm_1_1mpdf.html" title="Conditional probability density, e.g. modeling some dependencies.">mpdf</a>* <a class="code" href="classbdm_1_1enorm.html#baea4d49c657342b58297d68cda16d26" title="Return conditional density on the given RV, the remaining rvs will be in conditioning...">enorm&lt;sq_T&gt;::condition</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a> &amp;rvn )<span class="keyword"> const</span> 
    997 <a name="l01320"></a>01320 <span class="keyword">        </span>{ 
    998 <a name="l01321"></a>01321  
    999 <a name="l01322"></a>01322                 it_assert_debug ( <a class="code" href="classbdm_1_1epdf.html#c4b863ff84c7a4882fb3ad18556027f9" title="True if rv is assigned.">isnamed</a>(),<span class="stringliteral">"rvs are not assigned"</span> ); 
    1000 <a name="l01323"></a>01323  
    1001 <a name="l01324"></a>01324                 <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a> rvc = <a class="code" href="classbdm_1_1epdf.html#62c5b8ff71d9ebe6cd58d3c342eb1dc8" title="Description of the random variable.">rv</a>.<a class="code" href="classbdm_1_1RV.html#aec44dabdf0a6d90fbae95e1356eda39" title="Subtract another variable from the current one.">subt</a> ( rvn ); 
    1002 <a name="l01325"></a>01325                 it_assert_debug ( ( rvc._dsize() +rvn._dsize() ==<a class="code" href="classbdm_1_1epdf.html#62c5b8ff71d9ebe6cd58d3c342eb1dc8" title="Description of the random variable.">rv</a>._dsize() ),<span class="stringliteral">"wrong rvn"</span> ); 
    1003 <a name="l01326"></a>01326                 <span class="comment">//Permutation vector of the new R</span> 
    1004 <a name="l01327"></a>01327                 ivec irvn = rvn.<a class="code" href="classbdm_1_1RV.html#cbebdb5e0d30101a6eb63550ef701c55">dataind</a> ( <a class="code" href="classbdm_1_1epdf.html#62c5b8ff71d9ebe6cd58d3c342eb1dc8" title="Description of the random variable.">rv</a> ); 
    1005 <a name="l01328"></a>01328                 ivec irvc = rvc.<a class="code" href="classbdm_1_1RV.html#cbebdb5e0d30101a6eb63550ef701c55">dataind</a> ( <a class="code" href="classbdm_1_1epdf.html#62c5b8ff71d9ebe6cd58d3c342eb1dc8" title="Description of the random variable.">rv</a> ); 
    1006 <a name="l01329"></a>01329                 ivec perm=concat ( irvn , irvc ); 
    1007 <a name="l01330"></a>01330                 sq_T Rn ( <a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>,perm ); 
    1008 <a name="l01331"></a>01331  
    1009 <a name="l01332"></a>01332                 <span class="comment">//fixme - could this be done in general for all sq_T?</span> 
    1010 <a name="l01333"></a>01333                 mat S=Rn.to_mat(); 
    1011 <a name="l01334"></a>01334                 <span class="comment">//fixme</span> 
    1012 <a name="l01335"></a>01335                 <span class="keywordtype">int</span> n=rvn._dsize()-1; 
    1013 <a name="l01336"></a>01336                 <span class="keywordtype">int</span> end=<a class="code" href="classbdm_1_1enorm.html#2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>.rows()-1; 
    1014 <a name="l01337"></a>01337                 mat S11 = S.get ( 0,n, 0, n ); 
    1015 <a name="l01338"></a>01338                 mat S12 = S.get ( 0, n , rvn._dsize(), end ); 
    1016 <a name="l01339"></a>01339                 mat S22 = S.get ( rvn._dsize(), end, rvn._dsize(), end ); 
    1017 <a name="l01340"></a>01340  
    1018 <a name="l01341"></a>01341                 vec mu1 = <a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a> ( irvn ); 
    1019 <a name="l01342"></a>01342                 vec mu2 = <a class="code" href="classbdm_1_1enorm.html#c702a194720853570d08b65482f842c7" title="mean value">mu</a> ( irvc ); 
    1020 <a name="l01343"></a>01343                 mat A=S12*inv ( S22 ); 
    1021 <a name="l01344"></a>01344                 sq_T R_n ( S11 - A *S12.T() ); 
    1022 <a name="l01345"></a>01345  
    1023 <a name="l01346"></a>01346                 <a class="code" href="classbdm_1_1mlnorm.html" title="Normal distributed linear function with linear function of mean value;.">mlnorm&lt;sq_T&gt;</a>* tmp=<span class="keyword">new</span> <a class="code" href="classbdm_1_1mlnorm.html" title="Normal distributed linear function with linear function of mean value;.">mlnorm&lt;sq_T&gt;</a> ( ); 
    1024 <a name="l01347"></a>01347                 tmp-&gt;set_rv ( rvn ); tmp-&gt;set_rvc ( rvc ); 
    1025 <a name="l01348"></a>01348                 tmp-&gt;set_parameters ( A,mu1-A*mu2,R_n ); 
    1026 <a name="l01349"></a>01349                 <span class="keywordflow">return</span> tmp; 
    1027 <a name="l01350"></a>01350         } 
    1028 <a name="l01351"></a>01351  
    1029 <a name="l01353"></a>01353  
    1030 <a name="l01354"></a>01354         <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
    1031 <a name="l01355"></a>01355         std::ostream &amp;operator&lt;&lt; ( std::ostream &amp;os,  mlnorm&lt;sq_T&gt; &amp;ml ) 
    1032 <a name="l01356"></a>01356         { 
    1033 <a name="l01357"></a>01357                 os &lt;&lt; <span class="stringliteral">"A:"</span>&lt;&lt; ml.A&lt;&lt;endl; 
    1034 <a name="l01358"></a>01358                 os &lt;&lt; <span class="stringliteral">"mu:"</span>&lt;&lt; ml.mu_const&lt;&lt;endl; 
    1035 <a name="l01359"></a>01359                 os &lt;&lt; <span class="stringliteral">"R:"</span> &lt;&lt; ml.iepdf-&gt;_R().to_mat() &lt;&lt;endl; 
    1036 <a name="l01360"></a>01360                 <span class="keywordflow">return</span> os; 
    1037 <a name="l01361"></a>01361         }; 
    1038 <a name="l01362"></a>01362  
    1039 <a name="l01363"></a>01363 } 
    1040 <a name="l01364"></a>01364 <span class="preprocessor">#endif //EF_H</span> 
     971<a name="l01304"></a>01304 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     972<a name="l01305"></a><a class="code" href="classbdm_1_1enorm.html#adcb739443669ba0a8faf634196284c08">01305</a> <a class="code" href="classbdm_1_1shared__ptr.html">shared_ptr&lt;epdf&gt;</a> <a class="code" href="classbdm_1_1enorm.html#adcb739443669ba0a8faf634196284c08" title="Return marginal density on the given RV, the remainig rvs are intergrated out.">enorm&lt;sq_T&gt;::marginal</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a> &amp;rvn )<span class="keyword"> const</span> 
     973<a name="l01306"></a>01306 <span class="keyword"></span>{ 
     974<a name="l01307"></a>01307         <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm&lt;sq_T&gt;</a> *tmp = <span class="keyword">new</span> <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm&lt;sq_T&gt;</a> (); 
     975<a name="l01308"></a>01308         <a class="code" href="classbdm_1_1shared__ptr.html">shared_ptr&lt;epdf&gt;</a> narrow(tmp); 
     976<a name="l01309"></a>01309         <a class="code" href="classbdm_1_1enorm.html#adcb739443669ba0a8faf634196284c08" title="Return marginal density on the given RV, the remainig rvs are intergrated out.">marginal</a> ( rvn, *tmp ); 
     977<a name="l01310"></a>01310         <span class="keywordflow">return</span> narrow; 
     978<a name="l01311"></a>01311 } 
     979<a name="l01312"></a>01312  
     980<a name="l01313"></a>01313 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     981<a name="l01314"></a>01314 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1enorm.html#adcb739443669ba0a8faf634196284c08" title="Return marginal density on the given RV, the remainig rvs are intergrated out.">enorm&lt;sq_T&gt;::marginal</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a> &amp;rvn, <a class="code" href="classbdm_1_1enorm.html" title="Gaussian density with positive definite (decomposed) covariance matrix.">enorm&lt;sq_T&gt;</a> &amp;target )<span class="keyword"> const</span> 
     982<a name="l01315"></a>01315 <span class="keyword"></span>{ 
     983<a name="l01316"></a>01316         <a class="code" href="bdmerror_8h.html#a89a0f906b242b79c5d3d342291b2cab4" title="Throw std::runtime_exception if t is not true and NDEBUG is not defined.">bdm_assert_debug</a> (<a class="code" href="classbdm_1_1epdf.html#ac4b863ff84c7a4882fb3ad18556027f9" title="True if rv is assigned.">isnamed</a>(), <span class="stringliteral">&quot;rv description is not assigned&quot;</span>); 
     984<a name="l01317"></a>01317         ivec irvn = rvn.<a class="code" href="classbdm_1_1RV.html#acbebdb5e0d30101a6eb63550ef701c55">dataind</a> (<a class="code" href="classbdm_1_1epdf.html#a62c5b8ff71d9ebe6cd58d3c342eb1dc8" title="Description of the random variable.">rv</a>); 
     985<a name="l01318"></a>01318  
     986<a name="l01319"></a>01319         sq_T Rn (<a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>, irvn);  <span class="comment">// select rows and columns of R</span> 
     987<a name="l01320"></a>01320  
     988<a name="l01321"></a>01321         target.<a class="code" href="classbdm_1_1epdf.html#af423e28448dbb69ef4905295ec8de8ff" title="Name its rv.">set_rv</a> ( rvn ); 
     989<a name="l01322"></a>01322         target.set_parameters (<a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a> (irvn), Rn); 
     990<a name="l01323"></a>01323 } 
     991<a name="l01324"></a>01324  
     992<a name="l01325"></a>01325 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     993<a name="l01326"></a><a class="code" href="classbdm_1_1enorm.html#a40201ca12c6900f7aeb493bb7c582e20">01326</a> <a class="code" href="classbdm_1_1shared__ptr.html">shared_ptr&lt;mpdf&gt;</a> <a class="code" href="classbdm_1_1enorm.html#a40201ca12c6900f7aeb493bb7c582e20" title="Return conditional density on the given RV, the remaining rvs will be in conditioning...">enorm&lt;sq_T&gt;::condition</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a> &amp;rvn )<span class="keyword"> const</span> 
     994<a name="l01327"></a>01327 <span class="keyword"></span>{ 
     995<a name="l01328"></a>01328         <a class="code" href="classbdm_1_1mlnorm.html" title="Normal distributed linear function with linear function of mean value;.">mlnorm&lt;sq_T&gt;</a> *tmp = <span class="keyword">new</span> <a class="code" href="classbdm_1_1mlnorm.html" title="Normal distributed linear function with linear function of mean value;.">mlnorm&lt;sq_T&gt;</a> (); 
     996<a name="l01329"></a>01329         <a class="code" href="classbdm_1_1shared__ptr.html">shared_ptr&lt;mpdf&gt;</a> narrow(tmp); 
     997<a name="l01330"></a>01330         <a class="code" href="classbdm_1_1enorm.html#a40201ca12c6900f7aeb493bb7c582e20" title="Return conditional density on the given RV, the remaining rvs will be in conditioning...">condition</a> ( rvn, *tmp ); 
     998<a name="l01331"></a>01331         <span class="keywordflow">return</span> narrow; 
     999<a name="l01332"></a>01332 } 
     1000<a name="l01333"></a>01333  
     1001<a name="l01334"></a>01334 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     1002<a name="l01335"></a>01335 <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1enorm.html#a40201ca12c6900f7aeb493bb7c582e20" title="Return conditional density on the given RV, the remaining rvs will be in conditioning...">enorm&lt;sq_T&gt;::condition</a> ( <span class="keyword">const</span> <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a> &amp;rvn, <a class="code" href="classbdm_1_1mpdf.html" title="Conditional probability density, e.g. modeling , where  is random variable, rv, and...">mpdf</a> &amp;target )<span class="keyword"> const</span> 
     1003<a name="l01336"></a>01336 <span class="keyword"></span>{ 
     1004<a name="l01337"></a>01337         <span class="keyword">typedef</span> <a class="code" href="classbdm_1_1mlnorm.html" title="Normal distributed linear function with linear function of mean value;.">mlnorm&lt;sq_T&gt;</a> TMlnorm; 
     1005<a name="l01338"></a>01338  
     1006<a name="l01339"></a>01339         <a class="code" href="bdmerror_8h.html#a89a0f906b242b79c5d3d342291b2cab4" title="Throw std::runtime_exception if t is not true and NDEBUG is not defined.">bdm_assert_debug</a> (<a class="code" href="classbdm_1_1epdf.html#ac4b863ff84c7a4882fb3ad18556027f9" title="True if rv is assigned.">isnamed</a>(), <span class="stringliteral">&quot;rvs are not assigned&quot;</span>); 
     1007<a name="l01340"></a>01340         TMlnorm &amp;uptarget = <span class="keyword">dynamic_cast&lt;</span>TMlnorm &amp;<span class="keyword">&gt;</span>(target); 
     1008<a name="l01341"></a>01341  
     1009<a name="l01342"></a>01342         <a class="code" href="classbdm_1_1RV.html" title="Class representing variables, most often random variables.">RV</a> rvc = <a class="code" href="classbdm_1_1epdf.html#a62c5b8ff71d9ebe6cd58d3c342eb1dc8" title="Description of the random variable.">rv</a>.<a class="code" href="classbdm_1_1RV.html#aaec44dabdf0a6d90fbae95e1356eda39" title="Subtract another variable from the current one.">subt</a> (rvn); 
     1010<a name="l01343"></a>01343         <a class="code" href="bdmerror_8h.html#a89a0f906b242b79c5d3d342291b2cab4" title="Throw std::runtime_exception if t is not true and NDEBUG is not defined.">bdm_assert_debug</a> ( (rvc._dsize() + rvn._dsize() == <a class="code" href="classbdm_1_1epdf.html#a62c5b8ff71d9ebe6cd58d3c342eb1dc8" title="Description of the random variable.">rv</a>._dsize()), <span class="stringliteral">&quot;wrong rvn&quot;</span>); 
     1011<a name="l01344"></a>01344         <span class="comment">//Permutation vector of the new R</span> 
     1012<a name="l01345"></a>01345         ivec irvn = rvn.<a class="code" href="classbdm_1_1RV.html#acbebdb5e0d30101a6eb63550ef701c55">dataind</a> (<a class="code" href="classbdm_1_1epdf.html#a62c5b8ff71d9ebe6cd58d3c342eb1dc8" title="Description of the random variable.">rv</a>); 
     1013<a name="l01346"></a>01346         ivec irvc = rvc.<a class="code" href="classbdm_1_1RV.html#acbebdb5e0d30101a6eb63550ef701c55">dataind</a> (<a class="code" href="classbdm_1_1epdf.html#a62c5b8ff71d9ebe6cd58d3c342eb1dc8" title="Description of the random variable.">rv</a>); 
     1014<a name="l01347"></a>01347         ivec perm = concat (irvn , irvc); 
     1015<a name="l01348"></a>01348         sq_T Rn (<a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>, perm); 
     1016<a name="l01349"></a>01349  
     1017<a name="l01350"></a>01350         <span class="comment">//fixme - could this be done in general for all sq_T?</span> 
     1018<a name="l01351"></a>01351         mat S = Rn.to_mat(); 
     1019<a name="l01352"></a>01352         <span class="comment">//fixme</span> 
     1020<a name="l01353"></a>01353         <span class="keywordtype">int</span> n = rvn._dsize() - 1; 
     1021<a name="l01354"></a>01354         <span class="keywordtype">int</span> end = <a class="code" href="classbdm_1_1enorm.html#a2d92dde696b2a7a5b10ddef5d22ba2c2" title="Covariance matrix in decomposed form.">R</a>.rows() - 1; 
     1022<a name="l01355"></a>01355         mat S11 = S.get (0, n, 0, n); 
     1023<a name="l01356"></a>01356         mat S12 = S.get (0, n , rvn._dsize(), end); 
     1024<a name="l01357"></a>01357         mat S22 = S.get (rvn._dsize(), end, rvn._dsize(), end); 
     1025<a name="l01358"></a>01358  
     1026<a name="l01359"></a>01359         vec mu1 = <a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a> (irvn); 
     1027<a name="l01360"></a>01360         vec mu2 = <a class="code" href="classbdm_1_1enorm.html#ac702a194720853570d08b65482f842c7" title="mean value">mu</a> (irvc); 
     1028<a name="l01361"></a>01361         mat A = S12 * inv (S22); 
     1029<a name="l01362"></a>01362         sq_T R_n (S11 - A *S12.T()); 
     1030<a name="l01363"></a>01363  
     1031<a name="l01364"></a>01364         uptarget.set_rv (rvn); 
     1032<a name="l01365"></a>01365         uptarget.set_rvc (rvc); 
     1033<a name="l01366"></a>01366         uptarget.set_parameters (A, mu1 - A*mu2, R_n); 
     1034<a name="l01367"></a>01367 } 
     1035<a name="l01368"></a>01368  
     1036<a name="l01371"></a>01371 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     1037<a name="l01372"></a><a class="code" href="classbdm_1_1mgnorm.html#ab736332d20e418bf50d45836e129f339">01372</a> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgnorm.html#ab736332d20e418bf50d45836e129f339" title="set mean function">mgnorm&lt;sq_T &gt;::set_parameters</a> (<span class="keyword">const</span> <a class="code" href="classbdm_1_1shared__ptr.html">shared_ptr&lt;fnc&gt;</a> &amp;g0, <span class="keyword">const</span> sq_T &amp;R0) { 
     1038<a name="l01373"></a>01373         g = g0; 
     1039<a name="l01374"></a>01374         this-&gt;<a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>.set_parameters (zeros (g-&gt;dimension()), R0); 
     1040<a name="l01375"></a>01375 } 
     1041<a name="l01376"></a>01376  
     1042<a name="l01377"></a>01377 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     1043<a name="l01378"></a><a class="code" href="classbdm_1_1mgnorm.html#ab31d63472cf6a1030cd8dbd8094c1f6d">01378</a> <span class="keywordtype">void</span> <a class="code" href="classbdm_1_1mgnorm.html#ab31d63472cf6a1030cd8dbd8094c1f6d">mgnorm&lt;sq_T &gt;::condition</a> (<span class="keyword">const</span> vec &amp;cond) {this-&gt;<a class="code" href="classbdm_1_1mpdf__internal.html#a47bab632af15120c88aad647ae129468" title="Internal epdf used for sampling.">iepdf</a>._mu() = g-&gt;eval (cond);}; 
     1044<a name="l01379"></a>01379  
     1045<a name="l01381"></a>01381 <span class="keyword">template</span>&lt;<span class="keyword">class</span> sq_T&gt; 
     1046<a name="l01382"></a>01382 std::ostream &amp;operator&lt;&lt; (std::ostream &amp;os,  mlnorm&lt;sq_T&gt; &amp;ml) 
     1047<a name="l01383"></a>01383 { 
     1048<a name="l01384"></a>01384         os &lt;&lt; <span class="stringliteral">&quot;A:&quot;</span> &lt;&lt; ml.A &lt;&lt; endl; 
     1049<a name="l01385"></a>01385         os &lt;&lt; <span class="stringliteral">&quot;mu:&quot;</span> &lt;&lt; ml.mu_const &lt;&lt; endl; 
     1050<a name="l01386"></a>01386         os &lt;&lt; <span class="stringliteral">&quot;R:&quot;</span> &lt;&lt; ml._R() &lt;&lt; endl; 
     1051<a name="l01387"></a>01387         <span class="keywordflow">return</span> os; 
     1052<a name="l01388"></a>01388 }; 
     1053<a name="l01389"></a>01389  
     1054<a name="l01390"></a>01390 } 
     1055<a name="l01391"></a>01391 <span class="preprocessor">#endif //EF_H</span> 
    10411056</pre></div></div> 
    1042 <hr size="1"><address style="text-align: right;"><small>Generated on Wed Aug 5 00:06:47 2009 for mixpp by&nbsp; 
     1057<hr size="1"/><address style="text-align: right;"><small>Generated on Sun Aug 30 22:10:49 2009 for mixpp by&nbsp; 
    10431058<a href="http://www.doxygen.org/index.html"> 
    1044 <img src="doxygen.png" alt="doxygen" align="middle" border="0"></a> 1.5.9 </small></address> 
     1059<img class="footer" src="doxygen.png" alt="doxygen"/></a> 1.6.1 </small></address> 
    10451060</body> 
    10461061</html>