-#' EMGrank
+#' EMGrank
#'
#' Description de EMGrank
#'
#' @param maxi Nombre maximal d'itérations dans l'algorithme EM
#' @param X Régresseurs
#' @param Y Réponse
-#' @param tau Seuil pour accepter la convergence
+#' @param eps Seuil pour accepter la convergence
#' @param rank Vecteur des rangs possibles
#'
#' @return A list ...
#' LLF : log vraisemblance associé à cet échantillon, pour les valeurs estimées des paramètres
#'
#' @export
-EMGrank <- function(Pi, Rho, mini, maxi, X, Y, tau, rank, fast = TRUE)
+EMGrank <- function(Pi, Rho, mini, maxi, X, Y, eps, rank, fast = TRUE)
{
if (!fast)
{
# Function in R
- return(.EMGrank_R(Pi, Rho, mini, maxi, X, Y, tau, rank))
+ return(.EMGrank_R(Pi, Rho, mini, maxi, X, Y, eps, rank))
}
-
+
# Function in C
n <- nrow(X) #nombre d'echantillons
p <- ncol(X) #nombre de covariables
m <- ncol(Y) #taille de Y (multivarié)
k <- length(Pi) #nombre de composantes dans le mélange
- .Call("EMGrank", Pi, Rho, mini, maxi, X, Y, tau, rank, phi = double(p * m * k),
+ .Call("EMGrank", Pi, Rho, mini, maxi, X, Y, eps, rank, phi = double(p * m * k),
LLF = double(1), n, p, m, k, PACKAGE = "valse")
}
}
# R version - slow but easy to read
-.EMGrank_R <- function(Pi, Rho, mini, maxi, X, Y, tau, rank)
+.EMGrank_R <- function(Pi, Rho, mini, maxi, X, Y, eps, rank)
{
# matrix dimensions
- n <- dim(X)[1]
- p <- dim(X)[2]
- m <- dim(Rho)[2]
- k <- dim(Rho)[3]
-
+ n <- nrow(X)
+ p <- ncol(X)
+ m <- ncol(Y)
+ k <- length(Pi)
+
# init outputs
phi <- array(0, dim = c(p, m, k))
Z <- rep(1, n)
LLF <- 0
-
+
# local variables
Phi <- array(0, dim = c(p, m, k))
deltaPhi <- c()
# main loop
ite <- 1
- while (ite <= mini || (ite <= maxi && sumDeltaPhi > tau))
+ while (ite <= mini || (ite <= maxi && sumDeltaPhi > eps))
{
# M step: update for Beta ( and then phi)
for (r in 1:k)
{
- Z_indice <- seq_len(n)[Z == r] #indices where Z == r
+ Z_indice <- seq_len(n)[Z == r] #indices where Z == r
if (length(Z_indice) == 0)
next
# U,S,V = SVD of (t(Xr)Xr)^{-1} * t(Xr) * Yr
- s <- svd(MASS::ginv(crossprod(matricize(X[Z_indice, ]))) %*% crossprod(matricize(X[Z_indice,
- ]), matricize(Y[Z_indice, ])))
+ s <- svd(MASS::ginv(crossprod(matricize(X[Z_indice, ]))) %*%
+ crossprod(matricize(X[Z_indice, ]), matricize(Y[Z_indice, ])))
S <- s$d
# Set m-rank(r) singular values to zero, and recompose best rank(r) approximation
# of the initial product
S[(rank[r] + 1):length(S)] <- 0
phi[, , r] <- s$u %*% diag(S) %*% t(s$v) %*% Rho[, , r]
}
-
+
# Step E and computation of the loglikelihood
sumLogLLF2 <- 0
for (i in seq_len(n))
maxLogGamIR <- -Inf
for (r in seq_len(k))
{
- dotProduct <- tcrossprod(Y[i, ] %*% Rho[, , r] - X[i, ] %*% phi[,
- , r])
- logGamIR <- log(Pi[r]) + log(det(Rho[, , r])) - 0.5 * dotProduct
+ dotProduct <- tcrossprod(Y[i, ] %*% Rho[, , r] - X[i, ] %*% phi[, , r])
+ logGamIR <- log(Pi[r]) + log(gdet(Rho[, , r])) - 0.5 * dotProduct
# Z[i] = index of max (gam[i,])
if (logGamIR > maxLogGamIR)
{
}
sumLogLLF2 <- sumLogLLF2 + log(sumLLF1)
}
-
+
LLF <- -1/n * sumLogLLF2
-
+
# update distance parameter to check algorithm convergence (delta(phi, Phi))
- deltaPhi <- c(deltaPhi, max((abs(phi - Phi))/(1 + abs(phi)))) #TODO: explain?
+ deltaPhi <- c(deltaPhi, max((abs(phi - Phi))/(1 + abs(phi)))) #TODO: explain?
if (length(deltaPhi) > deltaPhiBufferSize)
deltaPhi <- deltaPhi[2:length(deltaPhi)]
sumDeltaPhi <- sum(abs(deltaPhi))
-
+
# update other local variables
Phi <- phi
ite <- ite + 1