install the package
[valse.git] / pkg / R / constructionModelesLassoMLE.R
... / ...
CommitLineData
1#' constructionModelesLassoMLE
2#'
3#' Construct a collection of models with the Lasso-MLE procedure.
4#'
5#' @param phiInit an initialization for phi, get by initSmallEM.R
6#' @param rhoInit an initialization for rho, get by initSmallEM.R
7#' @param piInit an initialization for pi, get by initSmallEM.R
8#' @param gamInit an initialization for gam, get by initSmallEM.R
9#' @param mini integer, minimum number of iterations in the EM algorithm, by default = 10
10#' @param maxi integer, maximum number of iterations in the EM algorithm, by default = 100
11#' @param gamma integer for the power in the penaly, by default = 1
12#' @param X matrix of covariates (of size n*p)
13#' @param Y matrix of responses (of size n*m)
14#' @param eps real, threshold to say the EM algorithm converges, by default = 1e-4
15#' @param S output of selectVariables.R
16#' @param ncores Number of cores, by default = 3
17#' @param fast TRUE to use compiled C code, FALSE for R code only
18#' @param verbose TRUE to show some execution traces
19#'
20#' @return a list with several models, defined by phi, rho, pi, llh
21#'
22#' @export
23constructionModelesLassoMLE <- function(phiInit, rhoInit, piInit, gamInit, mini,
24 maxi, gamma, X, Y, eps, S, ncores = 3, fast, verbose)
25{
26 if (ncores > 1)
27 {
28 cl <- parallel::makeCluster(ncores, outfile = "")
29 parallel::clusterExport(cl, envir = environment(), varlist = c("phiInit",
30 "rhoInit", "gamInit", "mini", "maxi", "gamma", "X", "Y", "eps", "S",
31 "ncores", "fast", "verbose"))
32 }
33
34 # Individual model computation
35 computeAtLambda <- function(lambda)
36 {
37 if (ncores > 1)
38 require("valse") #nodes start with an empty environment
39
40 if (verbose)
41 print(paste("Computations for lambda=", lambda))
42
43 n <- dim(X)[1]
44 p <- dim(phiInit)[1]
45 m <- dim(phiInit)[2]
46 k <- dim(phiInit)[3]
47 sel.lambda <- S[[lambda]]$selected
48 # col.sel = which(colSums(sel.lambda)!=0) #if boolean matrix
49 col.sel <- which(sapply(sel.lambda, length) > 0) #if list of selected vars
50 if (length(col.sel) == 0)
51 return(NULL)
52
53 # lambda == 0 because we compute the EMV: no penalization here
54 res <- EMGLLF(array(phiInit[col.sel, , ],dim=c(length(col.sel),m,k)), rhoInit,
55 piInit, gamInit, mini, maxi, gamma, 0, as.matrix(X[, col.sel]), Y, eps, fast)
56
57 # Eval dimension from the result + selected
58 phiLambda2 <- res$phi
59 rhoLambda <- res$rho
60 piLambda <- res$pi
61 phiLambda <- array(0, dim = c(p, m, k))
62 for (j in seq_along(col.sel))
63 phiLambda[col.sel[j], sel.lambda[[j]], ] <- phiLambda2[j, sel.lambda[[j]], ]
64 dimension <- length(unlist(sel.lambda))
65
66 ## Computation of the loglikelihood
67 # Precompute det(rhoLambda[,,r]) for r in 1...k
68 detRho <- sapply(1:k, function(r) det(rhoLambda[, , r]))
69 sumLogLLH <- 0
70 for (i in 1:n)
71 {
72 # Update gam[,]; use log to avoid numerical problems
73 logGam <- sapply(1:k, function(r) {
74 log(piLambda[r]) + log(detRho[r]) - 0.5 *
75 sum((Y[i, ] %*% rhoLambda[, , r] - X[i, ] %*% phiLambda[, , r])^2)
76 })
77
78 logGam <- logGam - max(logGam) #adjust without changing proportions
79 gam <- exp(logGam)
80 norm_fact <- sum(gam)
81 sumLogLLH <- sumLogLLH + log(norm_fact) - log((2 * base::pi)^(m/2))
82 }
83 llhLambda <- c(sumLogLLH/n, (dimension + m + 1) * k - 1)
84 # densite <- vector("double", n)
85 # for (r in 1:k)
86 # {
87 # if (length(col.sel) == 1)
88 # {
89 # delta <- (Y %*% rhoLambda[, , r] - (X[, col.sel] %*% t(phiLambda[col.sel, , r])))
90 # } else delta <- (Y %*% rhoLambda[, , r] - (X[, col.sel] %*% phiLambda[col.sel, , r]))
91 # densite <- densite + piLambda[r] * det(rhoLambda[, , r])/(sqrt(2 * base::pi))^m *
92 # exp(-rowSums(delta^2)/2)
93 # }
94 # llhLambda <- c(mean(log(densite)), (dimension + m + 1) * k - 1)
95 list(phi = phiLambda, rho = rhoLambda, pi = piLambda, llh = llhLambda)
96 }
97
98 # For each lambda, computation of the parameters
99 out <-
100 if (ncores > 1) {
101 parLapply(cl, 1:length(S), computeAtLambda)
102 } else {
103 lapply(1:length(S), computeAtLambda)
104 }
105
106 if (ncores > 1)
107 parallel::stopCluster(cl)
108
109 out
110}