1 | #include <stat/libEF.h> |
---|
2 | using namespace bdm; |
---|
3 | |
---|
4 | //These lines are needed for use of cout and endl |
---|
5 | using std::cout; |
---|
6 | using std::endl; |
---|
7 | |
---|
8 | void Test ( const egiw &E ) { |
---|
9 | } |
---|
10 | |
---|
11 | int main() { |
---|
12 | cout << "Testing eGiw(1,1)"<<endl; |
---|
13 | { |
---|
14 | // Setup model |
---|
15 | double mu=1.1; |
---|
16 | double s=0.1; |
---|
17 | |
---|
18 | // TEST 1x1 EGIW |
---|
19 | mat V ( 2,2 ); |
---|
20 | V ( 0,0 ) = pow ( mu,2 ) +s; |
---|
21 | V ( 1,0 ) = mu; |
---|
22 | V ( 0,1 ) = V ( 1,0 ); |
---|
23 | V ( 1,1 ) = 1.0; |
---|
24 | |
---|
25 | double nu=10; |
---|
26 | |
---|
27 | RV thr ( "{thr }", "2" ); |
---|
28 | egiw E ( thr,nu*V,nu ); |
---|
29 | cout << "egiw mean value:" << E.mean() <<endl; |
---|
30 | cout << "egiw normalizing constant:" << E.lognc() <<endl; |
---|
31 | |
---|
32 | int n=100; |
---|
33 | vec t_val ( 2 ); |
---|
34 | |
---|
35 | mat pdf ( 2*n,n ); |
---|
36 | vec Mu ( 2*n ); |
---|
37 | vec Si ( n ); |
---|
38 | |
---|
39 | for ( int i=0;i<2*n;i++ ) { |
---|
40 | Mu ( i ) = -2+i* ( 1.0/ ( n ) ) *3.0; |
---|
41 | t_val ( 0 ) = Mu ( i ); |
---|
42 | for ( int j=0;j<n;j++ ) { |
---|
43 | Si ( j ) = ( j+1 ) * ( 1.0/n ) *2; |
---|
44 | t_val ( 1 ) = Si ( j ); |
---|
45 | |
---|
46 | pdf ( i,j ) =E.evallog ( t_val ); |
---|
47 | } |
---|
48 | } |
---|
49 | |
---|
50 | mat Pdf=exp ( pdf ); |
---|
51 | vec fm=sum ( Pdf,2 ); |
---|
52 | vec fs=sum ( Pdf,1 ); |
---|
53 | cout << "Numerical mean: " << vec_2 ( Mu*fm/sum ( fm ), Si*fs/sum ( fs ) ) <<endl; |
---|
54 | cout << "Numerical integral of pdf: "<<sumsum ( Pdf/n/n*3*2 ) <<endl; |
---|
55 | } |
---|
56 | cout << "Testing Egiw(1,2)"<<endl; |
---|
57 | { |
---|
58 | // Setup model |
---|
59 | double mu=1.1; //unit step parametr |
---|
60 | double b=3.0; // sequence of <1 -1 1 -1...> |
---|
61 | double s=0.1; |
---|
62 | |
---|
63 | |
---|
64 | // TEST 1x1 EGIW |
---|
65 | mat V ( 3,3 ); |
---|
66 | V ( 0,0 ) = pow ( mu,2 ) +pow ( b ,2 ) +s; |
---|
67 | V ( 1,0 ) = mu; |
---|
68 | V ( 2,0 ) = b; |
---|
69 | |
---|
70 | V ( 0,1 ) = V ( 1,0 ); |
---|
71 | V ( 1,1 ) = 1.0; |
---|
72 | V ( 2,1 ) = 0.0; |
---|
73 | |
---|
74 | V ( 0,2 ) = V ( 2,0 ); |
---|
75 | V ( 1,2 ) = V ( 2,1 ); |
---|
76 | V ( 2,2 ) = 1.0; |
---|
77 | |
---|
78 | |
---|
79 | double nu=20; |
---|
80 | |
---|
81 | RV thr ( "{thr }", "3" ); |
---|
82 | egiw E ( thr,nu*V,nu ); |
---|
83 | cout << "egiw mean value:" << E.mean() <<endl; |
---|
84 | cout << "egiw normalizing constant:" << E.lognc() <<endl; |
---|
85 | |
---|
86 | int n=100; |
---|
87 | vec t_val ( 3 ); |
---|
88 | |
---|
89 | mat Tmp= zeros ( 2*n,n ); |
---|
90 | |
---|
91 | double summ=0.0; |
---|
92 | for ( int k=0;k<n;k++ ) { // ALL b |
---|
93 | t_val ( 1 ) = 1 + k* ( 1.0/n ) * 4.0; |
---|
94 | for ( int i=0;i<2*n;i++ ) { //ALL mu |
---|
95 | t_val ( 0 ) = -2+i* ( 1.0/ n ) *3.0; |
---|
96 | for ( int j=0;j<n;j++ ) { // All sigma |
---|
97 | t_val ( 2 ) = ( j+1 ) * ( 1.0/n ) *2.0; |
---|
98 | |
---|
99 | Tmp ( i,j ) = E.evallog ( t_val ); |
---|
100 | } |
---|
101 | } |
---|
102 | summ += sumsum ( exp ( Tmp ) ) /n/n/n*3.0*2.0*4.0; |
---|
103 | } |
---|
104 | |
---|
105 | |
---|
106 | cout << "Numerical integral of pdf: "<<summ <<endl; |
---|
107 | } |
---|
108 | |
---|
109 | |
---|
110 | } |
---|