test auto-indenter
[valse.git] / pkg / R / main.R
CommitLineData
086ca318
BA
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
43d76c49 15#' @param rank.min integer, minimum rank in the low rank procedure, by default = 1
16#' @param rank.max integer, maximum rank in the low rank procedure, by default = 5
aa480ac1
BA
17#' @param ncores_outer Number of cores for the outer loop on k
18#' @param ncores_inner Number of cores for the inner loop on lambda
43d76c49 19#' @param thresh real, threshold to say a variable is relevant, by default = 1e-8
aa480ac1
BA
20#' @param size_coll_mod (Maximum) size of a collection of models
21#' @param fast TRUE to use compiled C code, FALSE for R code only
22#' @param verbose TRUE to show some execution traces
086ca318
BA
23#'
24#' @return a list with estimators of parameters
25#'
26#' @examples
27#' #TODO: a few examples
28#' @export
2279a641 29valse = function(X, Y, procedure='LassoMLE', selecMod='DDSE', gamma=1, mini=10, maxi=50,
43d76c49 30 eps=1e-4, kmin=2, kmax=3, rank.min=1, rank.max=5, ncores_outer=1, ncores_inner=1,
31 thresh=1e-8,
fb6e49cb 32 size_coll_mod=10, fast=TRUE, verbose=FALSE, plot = TRUE)
086ca318 33{
086ca318
BA
34 p = dim(X)[2]
35 m = dim(Y)[2]
36 n = dim(X)[1]
fb6e49cb 37
4cc632c9 38 if (verbose)
fb6e49cb 39 print("main loop: over all k and all lambda")
40
41 if (ncores_outer > 1)
42 {
43 cl = parallel::makeCluster(ncores_outer, outfile='')
44 parallel::clusterExport( cl=cl, envir=environment(), varlist=c("X","Y","procedure",
43d76c49 45 "selecMod","gamma","mini","maxi","eps","kmin","kmax","rank.min","rank.max",
46 "ncores_outer","ncores_inner","thresh","size_coll_mod","verbose","p","m") )
fb6e49cb 47 }
48
49 # Compute models with k components
50 computeModels <- function(k)
51 {
52 if (ncores_outer > 1)
53 require("valse") #nodes start with an empty environment
54
55 if (verbose)
56 print(paste("Parameters initialization for k =",k))
57 #smallEM initializes parameters by k-means and regression model in each component,
086ca318
BA
58 #doing this 20 times, and keeping the values maximizing the likelihood after 10
59 #iterations of the EM algorithm.
4cc632c9
BA
60 P = initSmallEM(k, X, Y)
61 grid_lambda <- computeGridLambda(P$phiInit, P$rhoInit, P$piInit, P$gamInit, X, Y,
fb6e49cb 62 gamma, mini, maxi, eps, fast)
086cf723 63 if (length(grid_lambda)>size_coll_mod)
64 grid_lambda = grid_lambda[seq(1, length(grid_lambda), length.out = size_coll_mod)]
fb6e49cb 65
66 if (verbose)
67 print("Compute relevant parameters")
086ca318 68 #select variables according to each regularization parameter
0eb161e3
BA
69 #from the grid: S$selected corresponding to selected variables
70 S = selectVariables(P$phiInit, P$rhoInit, P$piInit, P$gamInit, mini, maxi, gamma,
43d76c49 71 grid_lambda, X, Y, thresh, eps, ncores_inner, fast)
086cf723 72
086ca318 73 if (procedure == 'LassoMLE')
fb6e49cb 74 {
4cc632c9 75 if (verbose)
fb6e49cb 76 print('run the procedure Lasso-MLE')
086ca318
BA
77 #compute parameter estimations, with the Maximum Likelihood
78 #Estimator, restricted on selected variables.
43d76c49 79 models <- constructionModelesLassoMLE( P$phiInit, P$rhoInit, P$piInit, P$gamInit,
80 mini, maxi, gamma, X, Y, eps, S, ncores_inner, fast, verbose)
81
086ca318 82 }
fb6e49cb 83 else
84 {
4cc632c9 85 if (verbose)
fb6e49cb 86 print('run the procedure Lasso-Rank')
086ca318
BA
87 #compute parameter estimations, with the Low Rank
88 #Estimator, restricted on selected variables.
43d76c49 89 models <- constructionModelesLassoRank(S, k, mini, maxi, X, Y, eps,
fb6e49cb 90 rank.min, rank.max, ncores_inner, fast, verbose)
086ca318 91 }
fb6e49cb 92 #warning! Some models are NULL after running selectVariables
93 models = models[sapply(models, function(cell) !is.null(cell))]
0eb161e3 94 models
086ca318 95 }
fb6e49cb 96
97 # List (index k) of lists (index lambda) of models
98 models_list <-
99 if (ncores_outer > 1)
100 parLapply(cl, kmin:kmax, computeModels)
101 else
102 lapply(kmin:kmax, computeModels)
103 if (ncores_outer > 1)
104 parallel::stopCluster(cl)
105
106 if (! requireNamespace("capushe", quietly=TRUE))
107 {
108 warning("'capushe' not available: returning all models")
109 return (models_list)
110 }
111
112 # Get summary "tableauRecap" from models
113 tableauRecap = do.call( rbind, lapply( seq_along(models_list), function(i) {
114 models <- models_list[[i]]
115 #For a collection of models (same k, several lambda):
116 LLH <- sapply( models, function(model) model$llh[1] )
117 k = length(models[[1]]$pi)
118 sumPen = sapply(models, function(model)
119 k*(dim(model$rho)[1]+sum(model$phi[,,1]!=0)+1)-1)
120 data.frame(model=paste(i,".",seq_along(models),sep=""),
121 pen=sumPen/n, complexity=sumPen, contrast=-LLH)
122 } ) )
123
124 print(tableauRecap)
125 tableauRecap = tableauRecap[which(tableauRecap[,4]!= Inf),]
fb6e49cb 126
7ac88d64 127 modSel = capushe::capushe(tableauRecap, n)
128 indModSel <-
129 if (selecMod == 'DDSE')
130 as.numeric(modSel@DDSE@model)
131 else if (selecMod == 'Djump')
132 as.numeric(modSel@Djump@model)
133 else if (selecMod == 'BIC')
134 modSel@BIC_capushe$model
135 else if (selecMod == 'AIC')
136 modSel@AIC_capushe$model
137
138 mod = as.character(tableauRecap[indModSel,1])
139 listMod = as.integer(unlist(strsplit(mod, "[.]")))
140 modelSel = models_list[[listMod[1]]][[listMod[2]]]
141
142 ##Affectations
143 Gam = matrix(0, ncol = length(modelSel$pi), nrow = n)
144 for (i in 1:n){
145 for (r in 1:length(modelSel$pi)){
146 sqNorm2 = sum( (Y[i,]%*%modelSel$rho[,,r]-X[i,]%*%modelSel$phi[,,r])^2 )
147 Gam[i,r] = modelSel$pi[r] * exp(-0.5*sqNorm2)* det(modelSel$rho[,,r])
148 }
149 }
150 Gam = Gam/rowSums(Gam)
151 modelSel$affec = apply(Gam, 1,which.max)
152 modelSel$proba = Gam
153
154 if (plot){
155 print(plot_valse(X,Y,modelSel,n))
156 }
157
158 return(modelSel)
086ca318 159}