move selectiontotale to appropriate folder (untranslated)
[valse.git] / src / sources / constructionModelesLassoMLE.c
CommitLineData
1d3c1faa 1#include "EMGLLF.h"
8e92c49c
BA
2#include "utils.h"
3#include <stdlib.h>
1d3c1faa
BA
4#include <gsl/gsl_linalg.h>
5#include <omp.h>
1d3c1faa
BA
6
7// TODO: comment on constructionModelesLassoMLE purpose
09ab3c16 8void constructionModelesLassoMLE_core(
3ec579a0 9 // IN parameters
9ff729fb
BA
10 const Real* phiInit, // parametre initial de moyenne renormalisé
11 const Real* rhoInit, // parametre initial de variance renormalisé
12 const Real* piInit,// parametre initial des proportions
13 const Real* gamInit, // paramètre initial des probabilités a posteriori de chaque échantillon
3ec579a0
BA
14 int mini,// nombre minimal d'itérations dans l'algorithme EM
15 int maxi,// nombre maximal d'itérations dans l'algorithme EM
46a2e676
BA
16 Real gamma,// valeur de gamma : puissance des proportions dans la pénalisation
17 //pour un Lasso adaptatif
9ff729fb
BA
18 const Real* glambda, // valeur des paramètres de régularisation du Lasso
19 const Real* X, // régresseurs
20 const Real* Y, // réponse
21 Real seuil,// seuil pour prendre en compte une variable
22 Real tau,// seuil pour accepter la convergence
3ec579a0
BA
23 const int* A1, // matrice des coefficients des parametres selectionnes
24 const int* A2, // matrice des coefficients des parametres non selectionnes
1d3c1faa 25 // OUT parameters
9ff729fb
BA
26 Real* phi,// estimateur ainsi calculé par le Lasso
27 Real* rho,// estimateur ainsi calculé par le Lasso
28 Real* pi, // estimateur ainsi calculé par le Lasso
c3bc4705 29 Real* llh, // estimateur ainsi calculé par le Lasso
1d3c1faa 30 // additional size parameters
3ec579a0
BA
31 int n, // taille de l'echantillon
32 int p, // nombre de covariables
33 int m, // taille de Y (multivarié)
34 int k, // nombre de composantes
35 int L) // taille de glambda
1d3c1faa 36{
46a2e676 37 //preparation: phi,rho,pi = 0, llh=+Inf
3ec579a0 38 for (int u=0; u<p*m*k*L; u++)
46a2e676
BA
39 phi[u] = 0.;
40 for (int u=0; u<m*m*k*L; u++)
41 rho[u] = 0.;
42 for (int u=0; u<k*L; u++)
43 pi[u] = 0.;
44 for (int u=0; u<L*2; u++)
45 llh[u] = INFINITY;
3ec579a0 46
1d3c1faa 47 //initiate parallel section
3ec579a0 48 int lambdaIndex;
1d3c1faa
BA
49 omp_set_num_threads(OMP_NUM_THREADS);
50 #pragma omp parallel default(shared) private(lambdaIndex)
51 {
52 #pragma omp for schedule(dynamic,CHUNK_SIZE) nowait
53 for (lambdaIndex=0; lambdaIndex<L; lambdaIndex++)
54 {
46a2e676 55 //a = A1[,1,lambdaIndex] ; a = a[a!=0]
3ec579a0
BA
56 int* a = (int*)malloc(p*sizeof(int));
57 int lengthA = 0;
58 for (int j=0; j<p; j++)
1d3c1faa 59 {
3ec579a0
BA
60 if (A1[ai(j,0,lambdaIndex,p,m+1,L)] != 0)
61 a[lengthA++] = A1[ai(j,0,lambdaIndex,p,m+1,L)] - 1;
1d3c1faa
BA
62 }
63 if (lengthA == 0)
46a2e676
BA
64 {
65 free(a);
1d3c1faa 66 continue;
46a2e676 67 }
3ec579a0 68
46a2e676 69 //Xa = X[,a]
9ff729fb 70 Real* Xa = (Real*)malloc(n*lengthA*sizeof(Real));
3ec579a0 71 for (int i=0; i<n; i++)
1d3c1faa 72 {
3ec579a0
BA
73 for (int j=0; j<lengthA; j++)
74 Xa[mi(i,j,n,lengthA)] = X[mi(i,a[j],n,p)];
1d3c1faa 75 }
3ec579a0 76
46a2e676 77 //phia = phiInit[a,,]
9ff729fb 78 Real* phia = (Real*)malloc(lengthA*m*k*sizeof(Real));
3ec579a0 79 for (int j=0; j<lengthA; j++)
1d3c1faa 80 {
3ec579a0 81 for (int mm=0; mm<m; mm++)
1d3c1faa 82 {
3ec579a0
BA
83 for (int r=0; r<k; r++)
84 phia[ai(j,mm,r,lengthA,m,k)] = phiInit[ai(a[j],mm,r,p,m,k)];
1d3c1faa
BA
85 }
86 }
3ec579a0 87
46a2e676 88 //Call to EMGLLF
9ff729fb
BA
89 Real* phiLambda = (Real*)malloc(lengthA*m*k*sizeof(Real));
90 Real* rhoLambda = (Real*)malloc(m*m*k*sizeof(Real));
91 Real* piLambda = (Real*)malloc(k*sizeof(Real));
92 Real* LLF = (Real*)malloc((maxi+1)*sizeof(Real));
93 Real* S = (Real*)malloc(lengthA*m*k*sizeof(Real));
46a2e676 94 EMGLLF_core(phia,rhoInit,piInit,gamInit,mini,maxi,gamma,0.,Xa,Y,tau,
1d3c1faa
BA
95 phiLambda,rhoLambda,piLambda,LLF,S,
96 n,lengthA,m,k);
97 free(Xa);
98 free(phia);
99 free(LLF);
100 free(S);
3ec579a0 101
46a2e676 102 //Assign results to current variables
3ec579a0 103 for (int j=0; j<lengthA; j++)
1d3c1faa 104 {
3ec579a0 105 for (int mm=0; mm<m; mm++)
1d3c1faa 106 {
3ec579a0 107 for (int r=0; r<k; r++)
46a2e676 108 phi[ai4(a[j],mm,r,lambdaIndex,p,m,k,L)] = phiLambda[ai(j,mm,r,lengthA,m,k)];
1d3c1faa
BA
109 }
110 }
111 free(phiLambda);
3ec579a0 112 for (int u=0; u<m; u++)
1d3c1faa 113 {
3ec579a0 114 for (int v=0; v<m; v++)
1d3c1faa 115 {
3ec579a0
BA
116 for (int r=0; r<k; r++)
117 rho[ai4(u,v,r,lambdaIndex,m,m,k,L)] = rhoLambda[ai(u,v,r,m,m,k)];
1d3c1faa
BA
118 }
119 }
120 free(rhoLambda);
3ec579a0
BA
121 for (int r=0; r<k; r++)
122 pi[mi(r,lambdaIndex,k,L)] = piLambda[r];
1d3c1faa 123 free(piLambda);
3ec579a0
BA
124
125 int dimension = 0;
126 int* b = (int*)malloc(m*sizeof(int));
127 for (int j=0; j<p; j++)
1d3c1faa 128 {
46a2e676 129 //b = A2[j,2:dim(A2)[2],lambdaIndex] ; b = b[b!=0]
3ec579a0
BA
130 int lengthB = 0;
131 for (int mm=0; mm<m; mm++)
1d3c1faa 132 {
3ec579a0
BA
133 if (A2[ai(j,mm+1,lambdaIndex,p,m+1,L)] != 0)
134 b[lengthB++] = A2[ai(j,mm+1,lambdaIndex,p,m+1,L)] - 1;
1d3c1faa 135 }
1d3c1faa
BA
136 if (lengthB > 0)
137 {
46a2e676 138 //phi[A2[j,1,lambdaIndex],b,,lambdaIndex] = 0.
3ec579a0 139 for (int mm=0; mm<lengthB; mm++)
1d3c1faa 140 {
3ec579a0 141 for (int r=0; r<k; r++)
46a2e676 142 phi[ai4(A2[ai(j,0,lambdaIndex,p,m+1,L)]-1, b[mm], r, lambdaIndex, p, m, k, L)] = 0.;
1d3c1faa
BA
143 }
144 }
3ec579a0 145
46a2e676 146 //c = A1[j,2:dim(A1)[2],lambdaIndex] ; dimension = dimension + sum(c!=0)
3ec579a0 147 for (int mm=0; mm<m; mm++)
1d3c1faa 148 {
3ec579a0 149 if (A1[ai(j,mm+1,lambdaIndex,p,m+1,L)] != 0)
1d3c1faa
BA
150 dimension++;
151 }
152 }
153 free(b);
3ec579a0 154
1d3c1faa 155 int signum;
9ff729fb
BA
156 Real* densite = (Real*)calloc(L*n,sizeof(Real));
157 Real sumLogDensit = 0.0;
1d3c1faa
BA
158 gsl_matrix* matrix = gsl_matrix_alloc(m, m);
159 gsl_permutation* permutation = gsl_permutation_alloc(m);
9ff729fb
BA
160 Real* YiRhoR = (Real*)malloc(m*sizeof(Real));
161 Real* XiPhiR = (Real*)malloc(m*sizeof(Real));
3ec579a0 162 for (int i=0; i<n; i++)
1d3c1faa 163 {
3ec579a0 164 for (int r=0; r<k; r++)
1d3c1faa
BA
165 {
166 //compute det(rho(:,:,r,lambdaIndex)) [TODO: avoid re-computations]
3ec579a0 167 for (int u=0; u<m; u++)
1d3c1faa 168 {
3ec579a0
BA
169 for (int v=0; v<m; v++)
170 matrix->data[u*m+v] = rho[ai4(u,v,r,lambdaIndex,m,m,k,L)];
1d3c1faa
BA
171 }
172 gsl_linalg_LU_decomp(matrix, permutation, &signum);
9ff729fb 173 Real detRhoR = gsl_linalg_LU_det(matrix, signum);
3ec579a0 174
1d3c1faa 175 //compute Y(i,:)*rho(:,:,r,lambdaIndex)
3ec579a0 176 for (int u=0; u<m; u++)
1d3c1faa
BA
177 {
178 YiRhoR[u] = 0.0;
3ec579a0
BA
179 for (int v=0; v<m; v++)
180 YiRhoR[u] += Y[mi(i,v,n,m)] * rho[ai4(v,u,r,lambdaIndex,m,m,k,L)];
1d3c1faa 181 }
3ec579a0 182
1d3c1faa 183 //compute X(i,a)*phi(a,:,r,lambdaIndex)
3ec579a0 184 for (int u=0; u<m; u++)
1d3c1faa
BA
185 {
186 XiPhiR[u] = 0.0;
3ec579a0
BA
187 for (int v=0; v<lengthA; v++)
188 XiPhiR[u] += X[mi(i,a[v],n,p)] * phi[ai4(a[v],u,r,lambdaIndex,p,m,k,L)];
1d3c1faa 189 }
46a2e676
BA
190 // NOTE: On peut remplacer X par Xa dans ce dernier calcul,
191 // mais je ne sais pas si c'est intéressant ...
3ec579a0 192
1d3c1faa 193 // compute dotProduct < delta . delta >
9ff729fb 194 Real dotProduct = 0.0;
3ec579a0 195 for (int u=0; u<m; u++)
1d3c1faa 196 dotProduct += (YiRhoR[u]-XiPhiR[u]) * (YiRhoR[u]-XiPhiR[u]);
3ec579a0 197
46a2e676
BA
198 densite[mi(lambdaIndex,i,L,n)] +=
199 (pi[mi(r,lambdaIndex,k,L)]*detRhoR/pow(sqrt(2.0*M_PI),m))*exp(-dotProduct/2.0);
3ec579a0 200 }
1d3c1faa
BA
201 sumLogDensit += log(densite[lambdaIndex*n+i]);
202 }
c3bc4705
BA
203 llh[mi(lambdaIndex,0,L,2)] = sumLogDensit;
204 llh[mi(lambdaIndex,1,L,2)] = (dimension+m+1)*k-1;
3ec579a0 205
1d3c1faa
BA
206 free(a);
207 free(YiRhoR);
208 free(XiPhiR);
209 free(densite);
210 gsl_matrix_free(matrix);
211 gsl_permutation_free(permutation);
212 }
213 }
214}