pkg/R/selectVariables.R

   1 #' selectVariables
   2 #'
   3 #' It is a function which constructs, for a given lambda, the sets for each cluster of relevant variables.
   4 #'
   5 #' @param phiInit an initial estimator for phi (size: p*m*k)
   6 #' @param rhoInit an initial estimator for rho (size: m*m*k)
   7 #' @param piInit an initial estimator for pi (size : k)
   8 #' @param gamInit an initial estimator for gamma
   9 #' @param mini  minimum number of iterations in EM algorithm
  10 #' @param maxi  maximum number of iterations in EM algorithm
  11 #' @param gamma  power in the penalty
  12 #' @param glambda grid of regularization parameters
  13 #' @param X    matrix of regressors
  14 #' @param Y    matrix of responses
  15 #' @param thresh real, threshold to say a variable is relevant, by default = 1e-8
  16 #' @param eps   threshold to say that EM algorithm has converged
  17 #' @param ncores Number or cores for parallel execution (1 to disable)
  18 #'
  19 #' @return a list of outputs, for each lambda in grid: selected,Rho,Pi
  20 #'
  21 #' @examples TODO
  22 #'
  23 #' @export
  24 #'
  25 selectVariables <- function(phiInit, rhoInit, piInit, gamInit, mini, maxi, gamma,
  26   glambda, X, Y, thresh = 1e-08, eps, ncores = 3, fast)
  27 {
  28   if (ncores > 1) {
  29     cl <- parallel::makeCluster(ncores, outfile = "")
  30     parallel::clusterExport(cl = cl, varlist = c("phiInit", "rhoInit", "gamInit",
  31       "mini", "maxi", "glambda", "X", "Y", "thresh", "eps"), envir = environment())
  32   }
  33
  34   # Computation for a fixed lambda
  35   computeCoefs <- function(lambda)
  36   {
  37     params <- EMGLLF(phiInit, rhoInit, piInit, gamInit, mini, maxi, gamma, lambda,
  38       X, Y, eps, fast)
  39
  40     p <- ncol(X)
  41     m <- ncol(Y)
  42
  43     # selectedVariables: list where element j contains vector of selected variables
  44     # in [1,m]
  45     selectedVariables <- lapply(1:p, function(j) {
  46       # from boolean matrix mxk of selected variables obtain the corresponding boolean
  47       # m-vector, and finally return the corresponding indices
  48       if (m>1) {
  49         seq_len(m)[apply(abs(params$phi[j, , ]) > thresh, 1, any)]
  50       } else {
  51         if (any(params$phi[j, , ] > thresh))
  52           1
  53         else
  54           numeric(0)
  55       }
  56     })
  57
  58     list(selected = selectedVariables, Rho = params$rho, Pi = params$pi)
  59   }
  60
  61   # For each lambda in the grid, we compute the coefficients
  62   out <-
  63     if (ncores > 1) {
  64       parLapply(cl, glambda, computeCoefs)
  65     } else {
  66       lapply(glambda, computeCoefs)
  67     }
  68   if (ncores > 1)
  69     parallel::stopCluster(cl)
  70
  71   print(out)
  72   # Suppress models which are computed twice En fait, ca ca fait la comparaison de
  73   # tous les parametres On veut juste supprimer ceux qui ont les memes variables
  74   # sélectionnées
  75   # sha1_array <- lapply(out, digest::sha1) out[ duplicated(sha1_array) ]
  76   selec <- lapply(out, function(model) model$selected)
  77   ind_dup <- duplicated(selec)
  78   ind_uniq <- which(!ind_dup)
  79   out2 <- list()
  80   for (l in 1:length(ind_uniq))
  81     out2[[l]] <- out[[ind_uniq[l]]]
  82   out2
  83 }