pkg/R/main.R

   1 #' valse
   2 #'
   3 #' Main function
   4 #'
   5 #' @param X matrix of covariates (of size n*p)
   6 #' @param Y matrix of responses (of size n*m)
   7 #' @param procedure among 'LassoMLE' or 'LassoRank'
   8 #' @param selecMod method to select a model among 'DDSE', 'DJump', 'BIC' or 'AIC'
   9 #' @param gamma integer for the power in the penaly, by default = 1
  10 #' @param mini integer, minimum number of iterations in the EM algorithm, by default = 10
  11 #' @param maxi integer, maximum number of iterations in the EM algorithm, by default = 100
  12 #' @param eps real, threshold to say the EM algorithm converges, by default = 1e-4
  13 #' @param kmin integer, minimum number of clusters, by default = 2
  14 #' @param kmax integer, maximum number of clusters, by default = 10
  15 #' @param rang.min integer, minimum rank in the low rank procedure, by default = 1
  16 #' @param rang.max integer, maximum rank in the
  17 #'
  18 #' @return a list with estimators of parameters
  19 #'
  20 #' @examples
  21 #' #TODO: a few examples
  22 #' @export
  23 valse = function(X, Y, procedure='LassoMLE', selecMod='DDSE', gamma=1, mini=10, maxi=50,
  24     eps=1e-4, kmin=2, kmax=2, rang.min=1, rang.max=10, ncores_outer=1, ncores_inner=3,
  25     verbose=FALSE)
  26 {
  27   p = dim(X)[2]
  28   m = dim(Y)[2]
  29   n = dim(X)[1]
  30
  31   if (verbose)
  32         print("main loop: over all k and all lambda")
  33
  34     if (ncores_outer > 1)
  35     {
  36         cl = parallel::makeCluster(ncores_outer)
  37         parallel::clusterExport( cl=cl, envir=environment(), varlist=c("X","Y","procedure",
  38             "selecMod","gamma","mini","maxi","eps","kmin","kmax","rang.min","rang.max",
  39             "ncores_outer","ncores_inner","verbose","p","m","k","tableauRecap") )
  40     }
  41
  42     # Compute models with k components
  43     computeModels <- function(k)
  44     {
  45         if (ncores_outer > 1)
  46             require("valse") #nodes start with an empty environment
  47
  48         if (verbose)
  49             print(paste("Parameters initialization for k =",k))
  50         #smallEM initializes parameters by k-means and regression model in each component,
  51     #doing this 20 times, and keeping the values maximizing the likelihood after 10
  52     #iterations of the EM algorithm.
  53     P = initSmallEM(k, X, Y)
  54     grid_lambda <- computeGridLambda(P$phiInit, P$rhoInit, P$piInit, P$gamInit, X, Y,
  55             gamma, mini, maxi, eps)
  56         # TODO: 100 = magic number
  57     if (length(grid_lambda)>100)
  58       grid_lambda = grid_lambda[seq(1, length(grid_lambda), length.out = 100)]
  59
  60         if (verbose)
  61             print("Compute relevant parameters")
  62     #select variables according to each regularization parameter
  63     #from the grid: S$selected corresponding to selected variables
  64     S = selectVariables(P$phiInit, P$rhoInit, P$piInit, P$gamInit, mini, maxi, gamma,
  65             grid_lambda, X, Y, 1e-8, eps, ncores_inner) #TODO: 1e-8 as arg?! eps?
  66
  67     if (procedure == 'LassoMLE')
  68         {
  69       if (verbose)
  70                 print('run the procedure Lasso-MLE')
  71       #compute parameter estimations, with the Maximum Likelihood
  72       #Estimator, restricted on selected variables.
  73       models <- constructionModelesLassoMLE(phiInit, rhoInit, piInit, gamInit, mini,
  74                 maxi, gamma, X, Y, thresh, eps, S$selected, ncores_inner, verbose)
  75     }
  76         else
  77         {
  78       if (verbose)
  79                 print('run the procedure Lasso-Rank')
  80       #compute parameter estimations, with the Low Rank
  81       #Estimator, restricted on selected variables.
  82       models <- constructionModelesLassoRank(S$Pi, S$Rho, mini, maxi, X, Y, eps, A1,
  83                 rank.min, rank.max, ncores_inner, verbose)
  84     }
  85     models
  86   }
  87
  88     # List (index k) of lists (index lambda) of models
  89     models_list <-
  90         if (ncores_k > 1)
  91             parLapply(cl, kmin:kmax, computeModels)
  92         else
  93             lapply(kmin:kmax, computeModels)
  94     if (ncores_k > 1)
  95         parallel::stopCluster(cl)
  96
  97     if (! requireNamespace("capushe", quietly=TRUE))
  98     {
  99         warning("'capushe' not available: returning all models")
 100         return (models_list)
 101     }
 102
 103     # Get summary "tableauRecap" from models ; TODO: jusqu'à ligne 114 à mon avis là c'est faux :/
 104     tableauRecap = t( sapply( models_list, function(models) {
 105         llh = do.call(rbind, lapply(models, function(model) model$llh)
 106     LLH = llh[-1,1]
 107     D = llh[-1,2]
 108         c(LLH, D, rep(k, length(model)), 1:length(model))
 109     ) } ) )
 110     if (verbose)
 111         print('Model selection')
 112   tableauRecap = tableauRecap[rowSums(tableauRecap[, 2:4])!=0,]
 113   tableauRecap = tableauRecap[!is.infinite(tableauRecap[,1]),]
 114   data = cbind(1:dim(tableauRecap)[1], tableauRecap[,2], tableauRecap[,2], tableauRecap[,1])
 115
 116   modSel = capushe::capushe(data, n)
 117   indModSel <-
 118         if (selecMod == 'DDSE')
 119             as.numeric(modSel@DDSE@model)
 120         else if (selecMod == 'Djump')
 121             as.numeric(modSel@Djump@model)
 122         else if (selecMod == 'BIC')
 123             modSel@BIC_capushe$model
 124         else if (selecMod == 'AIC')
 125             modSel@AIC_capushe$model
 126   model[[tableauRecap[indModSel,3]]][[tableauRecap[indModSel,4]]]
 127 }