pkg/R/selectVariables.R

   1 #' selectVariables
   2 #'
   3 #' For a given lambda, construct the sets of relevant variables for each cluster.
   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 #' @param fast boolean to enable or not the C function call
  19 #'
  20 #' @return a list, varying lambda in a grid, with selected (the indices of variables that are selected),
  21 #' Rho (the covariance parameter, reparametrized), Pi (the proportion parameter)
  22 #'
  23 #' @export
  24 selectVariables <- function(phiInit, rhoInit, piInit, gamInit, mini, maxi, gamma,
  25   glambda, X, Y, thresh = 1e-08, eps, ncores = 3, fast)
  26 {
  27   if (ncores > 1) {
  28     cl <- parallel::makeCluster(ncores, outfile = "")
  29     parallel::clusterExport(cl = cl, varlist = c("phiInit", "rhoInit", "gamInit",
  30       "mini", "maxi", "glambda", "X", "Y", "thresh", "eps"), envir = environment())
  31   }
  32
  33   # Computation for a fixed lambda
  34   computeCoefs <- function(lambda)
  35   {
  36     params <- EMGLLF(phiInit, rhoInit, piInit, gamInit, mini, maxi, gamma, lambda,
  37       X, Y, eps, fast)
  38
  39     p <- ncol(X)
  40     m <- ncol(Y)
  41
  42     # selectedVariables: list where element j contains vector of selected variables
  43     # in [1,m]
  44     selectedVariables <- lapply(1:p, function(j) {
  45       # from boolean matrix mxk of selected variables obtain the corresponding boolean
  46       # m-vector, and finally return the corresponding indices
  47       if (m>1) {
  48         seq_len(m)[apply(abs(params$phi[j, , ]) > thresh, 1, any)]
  49       } else {
  50         if (any(params$phi[j, , ] > thresh))
  51           1
  52         else
  53           numeric(0)
  54       }
  55     })
  56
  57     list(selected = selectedVariables, Rho = params$rho, Pi = params$pi)
  58   }
  59
  60   # For each lambda in the grid, we compute the coefficients
  61   out <-
  62     if (ncores > 1) {
  63       parLapply(cl, glambda, computeCoefs)
  64     } else {
  65       lapply(glambda, computeCoefs)
  66     }
  67   if (ncores > 1)
  68     parallel::stopCluster(cl)
  69
  70   # Suppress models which are computed twice
  71   # sha1_array <- lapply(out, digest::sha1) out[ duplicated(sha1_array) ]
  72   selec <- lapply(out, function(model) model$selected)
  73   ind_dup <- duplicated(selec)
  74   ind_uniq <- which(!ind_dup)
  75   out2 <- list()
  76   for (l in 1:length(ind_uniq))
  77     out2[[l]] <- out[[ind_uniq[l]]]
  78   out2
  79 }