mixpp: bdm::PF Class Reference
bdm::PF

bdm::PF Class Reference

Trivial particle filter with proposal density equal to parameter evolution model. More...

#include <particles.h>

Inheritance diagram for bdm::PF:

bdm::BM bdm::root List of all members.

Public Types

 __VA_ARGS__
enum  log_level_enums { __VA_ARGS__ }

Public Member Functions

virtual void bayes_weights ()
 bayes compute weights of the
virtual bool do_resampling ()
 important part of particle filtering - decide if it is time to perform resampling
void bayes (const vec &yt, const vec &cond)
vec & _lls ()
 access function
RESAMPLING_METHOD _resmethod () const
 access function
const pf_mix & posterior () const
 return correctly typed posterior (covariant return)
void from_setting (const Setting &set)
void to_setting (Setting &set) const
void log_register (bdm::logger &L, const string &prefix)
void log_write () const
void set_prior (const epdf *pri)
void resmethod_from_set (const Setting &set)
 auxiliary function reading parameter 'resmethod' from configuration file
void validate ()
void resample ()
 resample posterior density (from outside - see MPF)
Array< BM * > & _particles ()
 access function
Constructors
void set_parameters (int n0, double res_th0=0.5, RESAMPLING_METHOD rm=SYSTEMATIC)
void set_model (const BM *particle0, const epdf *prior)
void set_statistics (const vec w0, const epdf &epdf0)

Public Attributes

log_level_template< PF > log_level

Protected Attributes

int n
 number of particles;
pf_mix est
 posterior density
vec w
 weights;
Array< BM * > particles
 particles
vec lls
 internal structure storing loglikelihood of predictions
RESAMPLING_METHOD resmethod
 which resampling method will be used
double res_threshold

Friends

class log_level_intermediate< PF >

Classes

class  pf_mix

Detailed Description

Trivial particle filter with proposal density equal to parameter evolution model.

Posterior density is represented by a weighted empirical density (eEmp ).

Member Function Documentation

void bdm::PF::from_setting ( const Setting &  set  )  [inline, virtual]

configuration structure for basic PF 

    particle        = bdm::BootstrapParticle;       % one bayes rule for each point in the empirical support
      - or -        = bdm::MarginalizedParticle;    % (in case of Marginalized Particle filtering
    prior           = epdf_class;                   % prior probability density on the empirical variable
    --- optional ---
    n               = 10;                           % number of particles
    resmethod       = 'systematic', or 'multinomial', or 'stratified'
                                                    % resampling method
    res_threshold   = 0.5;                          % resample when active particles drop below 50%

Reimplemented from bdm::BM.

Member Data Documentation

double bdm::PF::res_threshold [protected]

resampling threshold; in this case its meaning is minimum ratio of active particles For example, for 0.5 resampling is performed when the numebr of active aprticles drops belo 50%.

The documentation for this class was generated from the following files:
Generated on 2 Dec 2013 for mixpp by  doxygen 1.4.7