added a class to the output and documented the function

DESCRIPTION

@@ -19,4 +19,4 @@ Imports:
     RSpectra,
     glmnet
 License: GPL (>= 2)
-RoxygenNote: 7.1.1
+RoxygenNote: 7.1.2

NAMESPACE

+# Generated by roxygen2: do not edit by hand
+
+S3method(print,SISIRres)
+S3method(print,ridgeRes)
+S3method(print,sparseRes)
+S3method(project,sparseRes)
+S3method(summary,SISIRres)
+S3method(summary,ridgeRes)
+S3method(summary,sparseRes)
+export(SISIR)
+export(project)
+export(ridgeSIR)
+export(sfcb)
+export(sparseSIR)
+export(tune.ridgeSIR)
+import(adjclust)
+import(doParallel)
+import(foreach)
+import(glmnet)
+import(magrittr)
+importFrom(Boruta,Boruta)
+importFrom(CORElearn,attrEval)
+importFrom(Matrix,forceSymmetric)
+importFrom(RSpectra,eigs)
+importFrom(dplyr,arrange)
+importFrom(dplyr,filter)
+importFrom(dplyr,inner_join)
+importFrom(dplyr,mutate)
+importFrom(dplyr,select)
+importFrom(dplyr,summarise)
+importFrom(dplyr,ungroup)
+importFrom(expm,sqrtm)
+importFrom(mixOmics,pls)
+importFrom(parallel,detectCores)
+importFrom(purrr,map_dfc)
+importFrom(ranger,ranger)
+importFrom(stats,cor)
+importFrom(stats,cov)
+importFrom(stats,predict)
+importFrom(stats,quantile)
+importFrom(tidyr,gather)
+importFrom(tidyr,spread)
+importFrom(utils,head)

R/SFCB.R

-## Q: name of the main function?
-
-# Main function ####
-# inputs: 
-# parallel : à voir (+ uniformiser dans SISIR)
-# rep : répétitions de l'algorithme RF
-# Q: which choices for rep?
-# Q: other inputs?
-#
-# outputs:
-# points sélectionnés ou importance + statistiques sur les "vrais" intervalles (mieux a posteriori) ?
-# mse
-# paramètres d'entrées
-# computational times
-# Q: other outputs?
-
+################################################################################
+# SFCB
+################################################################################
+#' @title SFCB
 #' @export
+#'
+#' @description
+#' \code{SFCB} performs interval selection based on random forests
+#'
+#' @param X input predictors (matrix or data.frame)
+#' @param Y target variable (vector whose length is equal to the number of rows
+#' in X)
+#' @param group.method group method. Default to \code{"adjclust"}
+#' @param summary.method summary method. Default to \code{"pls"}
+#' @param selection.method selection method. Default to \code{"none"} (no 
+#' selection performed)
+#' @param at number of groups targeted for output results (integer). Not used
+#' when \code{range.at} is not \code{NULL}
+#' @param range.at (vector of integer) sequence of the numbers of groups for
+#' output results
+#' @param seed random seed (integer)
+#' @param repeats number of repeats for the final random forest computation
+#' @param parallel not implemented yet
+#' 
+#' @return an object of class \code{"SFCB"} with elements: 
+#'   \item{dendro}{a dendrogram corresponding to the method chosen in
+#'   \code{group.method}}
+#'   \item{groups}{a list of length \code{length(range.at)} (or of length 1 if
+#'   \code{range.at == NULL}) that contains the clusterings of input variables
+#'   for the selected group numbers}
+#'   \item{summaries}{a list of the same length than \code{$groups} that 
+#'   contains the summarized predictors according to the method chosen in 
+#'   \code{summary.methods}}
+#'   \item{selected}{a list of the same length than \code{$groups} that contains
+#'   the names of the variable selected by \code{selection.method} if it is not
+#'   equal to \code{"none"}}
+#'   \item{mse}{a data.frame with \code{repeats} \eqn{\times}
+#'   \code{length($groups)} rows that contains Mean Squared Errors of the 
+#'   \code{repeats} random forests fitted for each number of groups}
+#'   \item{importance}{a list of the same length than \code{$groups} that 
+#'   contains a data.frame providing variable importances for the variables in
+#'   selected groups in \code{repeats} columns (one for each iteration of the
+#'   random forest method). When \code{summary.method == "basics"}, importance 
+#'   for mean and sd are provided in separated columns, in which case, the 
+#'   number of columns is equal to 2\code{repeats}}
+#' 
+#' @author {Remi Servien, \email{remi.servien@inrae.fr}\cr
+#' Nathalie Vialaneix, \email{nathalie.vialaneix@inrae.fr}}
+#' 
+## TODO: add seealso, details, references and example in doc
+## TODO: add parallel backend
+## TODO: add computational time computations
 
 sfcb <- function(X, Y, group.method = c("adjclust", "cclustofvar"), 
                  summary.method = c("pls", "basics", "cclustofvar"), 
                  selection.method = c("none", "boruta", "relief"), 
                  at = round(0.15 * ncol(X)), range.at = NULL,  seed = NULL, 
                  repeats = 5, parallel = FALSE) {
+  
   # Input checking ####
   group.method <- match.arg(group.method)
   summary.method <- match.arg(summary.method)
@@ -131,6 +167,8 @@ sfcb <- function(X, Y, group.method = c("adjclust", "cclustofvar"),
                 "importances" = all_qualities$importances)
   }
   
+  class(out) <- "SFCB"
+  
   return(out)
 }
 
@@ -147,7 +185,6 @@ sfcb <- function(X, Y, group.method = c("adjclust", "cclustofvar"),
 # out9 <- sfcb(x, Y, group.method = "adjclust", summary.method = "basics", selection.method = "relief")
 # out10 <- sfcb(x, Y, group.method = "adjclust", summary.method = "pls", selection.method = "boruta", seed = 3)
 
-
 # Methods #####
 # plots
 # pour "at"
@@ -155,4 +192,4 @@ sfcb <- function(X, Y, group.method = c("adjclust", "cclustofvar"),
 # pour range_at représenter l'évolution de la mse + représenter une dizaine d'intervalles max
 # aussi : option dendrogramme pour représenter le dendrogramme avec le at ou le range_at identifié dessus
 # qualities (requiert la vérité) + un éventuel seuil pour l'importance et calcule les valeurs de qualité (rand index, différence entre les importances avec un boxplot ?)
-#   ajout d'une méthode de plot dessus pour range_at, pour voir l'évolution de la qualité
+#   ajout d'une méthode de plot dessus pour range_at, pour voir l'évolution de la qualité
\ No newline at end of file

man/SISIR.Rd

+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/sparseSIR.R
+\name{SISIR}
+\alias{SISIR}
+\title{Interval Sparse SIR}
+\usage{
+SISIR(
+  object,
+  inter_len = rep(1, nrow(object$EDR)),
+  sel_prop = 0.05,
+  itermax = Inf,
+  minint = 2,
+  parallel = TRUE,
+  ncores = NULL
+)
+}
+\arguments{
+\item{object}{an object of class \code{ridgeRes} as obtained from the 
+function \code{\link{ridgeSIR}}}
+
+\item{inter_len}{(numeric) vector with interval lengths for the initial 
+state. Default is to set one interval for each variable (all intervals have
+length 1)}
+
+\item{sel_prop}{fraction of the coefficients that will be considered as 
+strong zeros and strong non zeros. Default to 0.05}
+
+\item{itermax}{maximum number of iterations. Default to Inf}
+
+\item{minint}{minimum number of intervals. Default to 2}
+
+\item{parallel}{whether the computation should be performed in parallel or
+not. Logical. Default is FALSE}
+
+\item{ncores}{number of cores to use if \code{parallel = TRUE}. If left to 
+NULL, all available cores minus one are used}
+}
+\value{
+S3 object of class \code{SISIR}: a list consisting of
+\itemize{
+   \item{\code{sEDR}}{ the estimated EDR spaces (a list of p x d matrices)}
+   \item{\code{alpha}}{ the estimated shrinkage coefficients (a list of 
+   vectors)}
+   \item{\code{intervals}}{ the interval lengths (a list of vectors)}
+   \item{\code{quality}}{ a data frame with various qualities for the model.
+   The chosen quality measures are the same than for the function 
+   \code{\link{sparseSIR}} plus the number of intervals \code{nbint}}
+   \item{\code{init_sel_prop}}{ initial fraction of the coefficients which 
+   are considered as strong zeros or strong non zeros}
+   \item{\code{rSIR}}{ same as the input \code{object}}
+ }
+}
+\description{
+\code{SISIR} performs an automatic search of relevant intervals
+}
+\details{
+Different quality criteria used to select the best models among a list of 
+models with different interval definitions. Quality criteria are: 
+log-likelihood (\code{loglik}), cross-validation error as provided by the
+function \code{\link[glmnet]{glmnet}}, two versions of the AIC (\code{AIC} 
+and \code{AIC2}) and of the BIC (\code{BIC} and \code{BIC2}) in which the 
+number of parameters is either the number of non null intervals or the 
+number of non null parameters with respect to the original variables
+}
+\examples{
+set.seed(1140)
+tsteps <- seq(0, 1, length = 200)
+nsim <- 100
+simulate_bm <- function() return(c(0, cumsum(rnorm(length(tsteps)-1, sd=1))))
+x <- t(replicate(nsim, simulate_bm()))
+beta <- cbind(sin(tsteps*3*pi/2), sin(tsteps*5*pi/2))
+beta[((tsteps < 0.2) | (tsteps > 0.5)), 1] <- 0
+beta[((tsteps < 0.6) | (tsteps > 0.75)), 2] <- 0
+y <- log(abs(x \%*\% beta[ ,1]) + 1) + sqrt(abs(x \%*\% beta[ ,2]))
+y <- y + rnorm(nsim, sd = 0.1)
+res_ridge <- ridgeSIR(x, y, H = 10, d = 2, mu2 = 10^8)
+\dontrun{res_fused <- SISIR(res_ridge, rep(1, ncol(x)))}
+
+}
+\references{
+{Picheny, V., Servien, R. and Villa-Vialaneix, N. (2016) 
+Interpretable sparse SIR for digitized functional data.
+\emph{Statistics and Computing}, \strong{29}(2), 255--267.}
+}
+\seealso{
+\code{\link{ridgeSIR}}, \code{\link{sparseSIR}}
+}
+\author{
+{Victor Picheny, \email{victor.picheny@inrae.fr}\cr
+Remi Servien, \email{remi.servien@inrae.fr}\cr
+Nathalie Vialaneix, \email{nathalie.vialaneix@inrae.fr}}
+}

man/SISIRres.Rd

+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/sparseSIR.R
+\name{SISIRres}
+\alias{SISIRres}
+\alias{summary.SISIRres}
+\alias{print.SISIRres}
+\title{Print SISIRres object}
+\usage{
+\method{summary}{SISIRres}(object, ...)
+
+\method{print}{SISIRres}(x, ...)
+}
+\arguments{
+\item{object}{a \code{SISIRres} object}
+
+\item{...}{not used}
+
+\item{x}{a \code{SISIRres} object}
+}
+\description{
+Print a summary of the result of \code{\link{SISIRres}} (
+\code{SISIRres} object)
+}
+\seealso{
+\code{\link{SISIR}}
+}
+\author{
+{Victor Picheny, \email{victor.picheny@inrae.fr}\cr
+Remi Servien, \email{remi.servien@inrae.fr}\cr
+Nathalie Vialaneix, \email{nathalie.vialaneix@inrae.fr}}
+}

man/project.Rd

+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/sparseSIR.R
+\name{project}
+\alias{project}
+\alias{project.sparseRes}
+\title{sparse SIR}
+\usage{
+\method{project}{sparseRes}(object)
+
+project(object)
+}
+\arguments{
+\item{object}{an object of class \code{sparseRes} as obtained from the 
+function \code{\link{sparseSIR}}}
+}
+\value{
+a matrix of dimension n x d with the projection of the observations
+on the d dimensions of the sparse EDR space
+}
+\description{
+\code{project} performs the projection on the sparse EDR space (as obtained
+by the \code{\link[glmnet]{glmnet}})
+}
+\details{
+The projection is obtained by the function 
+\code{\link[glmnet]{predict.glmnet}}.
+}
+\examples{
+set.seed(1140)
+tsteps <- seq(0, 1, length = 200)
+nsim <- 100
+simulate_bm <- function() return(c(0, cumsum(rnorm(length(tsteps)-1, sd=1))))
+x <- t(replicate(nsim, simulate_bm()))
+beta <- cbind(sin(tsteps*3*pi/2), sin(tsteps*5*pi/2))
+beta[((tsteps < 0.2) | (tsteps > 0.5)), 1] <- 0
+beta[((tsteps < 0.6) | (tsteps > 0.75)), 2] <- 0
+y <- log(abs(x \%*\% beta[ ,1]) + 1) + sqrt(abs(x \%*\% beta[ ,2]))
+y <- y + rnorm(nsim, sd = 0.1)
+\dontrun{
+res_ridge <- ridgeSIR(x, y, H = 10, d = 2)
+res_sparse <- sparseSIR(res_ridge, rep(1, ncol(x)))
+proj_data <- project(res_sparse)
+}
+
+}
+\references{
+{Picheny, V., Servien, R. and Villa-Vialaneix, N. (2016) 
+Interpretable sparse SIR for digitized functional data.
+\emph{Statistics and Computing}, \strong{29}(2), 255--267.}
+}
+\seealso{
+\code{\link{sparseSIR}}
+}
+\author{
+{Victor Picheny, \email{victor.picheny@inrae.fr}\cr
+Remi Servien, \email{remi.servien@inrae.fr}\cr
+Nathalie Vialaneix, \email{nathalie.vialaneix@inrae.fr}}
+}

man/ridgeRes.Rd

+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/ridgeSIR.R
+\name{ridgeRes}
+\alias{ridgeRes}
+\alias{summary.ridgeRes}
+\alias{summary.ridgeRes()}
+\alias{print.ridgeRes}
+\alias{ridgeRes-class}
+\title{Print ridgeRes object}
+\usage{
+\method{summary}{ridgeRes}(object, ...)
+
+\method{print}{ridgeRes}(x, ...)
+}
+\arguments{
+\item{object}{a \code{ridgeRes} object}
+
+\item{...}{not used}
+
+\item{x}{a \code{ridgeRes} object}
+}
+\description{
+Print a summary of the result of \code{\link{ridgeSIR}} (
+\code{ridgeRes} object)
+}
+\seealso{
+\code{\link{ridgeSIR}}
+}
+\author{
+{Victor Picheny, \email{victor.picheny@inrae.fr}\cr 
+Remi Servien, \email{remi.servien@inrae.fr}\cr
+Nathalie Vialaneix, \email{nathalie.vialaneix@inrae.fr}}
+}

man/ridgeSIR.Rd

+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/ridgeSIR.R
+\name{ridgeSIR}
+\alias{ridgeSIR}
+\title{ridge SIR}
+\usage{
+ridgeSIR(x, y, H, d, mu2 = NULL)
+}
+\arguments{
+\item{x}{explanatory variables (numeric matrix or data frame)}
+
+\item{y}{target variable (numeric vector)}
+
+\item{H}{number of slices (integer)}
+
+\item{d}{number of dimensions to be kept}
+
+\item{mu2}{ridge regularization parameter (numeric, positive)}
+}
+\value{
+S3 object of class \code{ridgeRes}: a list consisting of
+\itemize{
+   \item{\code{EDR}}{ the estimated EDR space (a p x d matrix)}
+   \item{\code{condC}}{ the estimated slice projection on EDR (a d x H 
+   matrix)}
+   \item{\code{eigenvalues}}{ the eigenvalues obtained during the generalized 
+   eigendecomposition performed by SIR}
+   \item{\code{parameters}}{ a list of hyper-parameters for the method: 
+   \itemize{
+     \item{\code{H}}{ number of slices}
+     \item{\code{d}}{ dimension of the EDR space}
+     \item{\code{mu2}}{ regularization parameter for the ridge penalty}
+   }}
+   \item{\code{utils}}{ useful outputs for further computations:
+   \itemize{
+     \item{\code{Sigma}}{ covariance matrix for x}
+     \item{\code{slices}}{ slice number for all observations}
+     \item{\code{invsqrtS}}{ value of the inverse square root of the 
+     regularized covariance matrix for x}
+   }}
+ }
+}
+\description{
+\code{ridgeSIR} performs the first step of the method (ridge regularization
+of SIR)
+}
+\details{
+SI-SIR
+}
+\examples{
+set.seed(1140)
+tsteps <- seq(0, 1, length = 50)
+simulate_bm <- function() return(c(0, cumsum(rnorm(length(tsteps)-1, sd=1))))
+x <- t(replicate(50, simulate_bm()))
+beta <- cbind(sin(tsteps*3*pi/2), sin(tsteps*5*pi/2)) 
+y <- log(abs(x \%*\% beta[ ,1])) + sqrt(abs(x \%*\% beta[ ,2]))
+y <- y + rnorm(50, sd = 0.1)
+res_ridge <- ridgeSIR(x, y, H = 10, d = 2, mu2 = 10^8)
+\dontrun{print(res_ridge)}
+
+}
+\references{
+{Picheny, V., Servien, R. and Villa-Vialaneix, N. (2016) 
+Interpretable sparse SIR for digitized functional data. 
+\emph{Statistics and Computing}, \strong{29}(2), 255--267.}
+}
+\seealso{
+\code{\link{sparseSIR}}, \code{\link{SISIR}}, 
+\code{\link{tune.ridgeSIR}}
+}
+\author{
+{Victor Picheny, \email{victor.picheny@inrae.fr}\cr
+Remi Servien, \email{remi.servien@inrae.fr}\cr
+Nathalie Vialaneix, \email{nathalie.vialaneix@inrae.fr}}
+}

man/sfcb.Rd

+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/SFCB.R
+\name{sfcb}
+\alias{sfcb}
+\title{SFCB}
+\usage{
+sfcb(
+  X,
+  Y,
+  group.method = c("adjclust", "cclustofvar"),
+  summary.method = c("pls", "basics", "cclustofvar"),
+  selection.method = c("none", "boruta", "relief"),
+  at = round(0.15 * ncol(X)),
+  range.at = NULL,
+  seed = NULL,
+  repeats = 5,
+  parallel = FALSE
+)
+}
+\arguments{
+\item{X}{input predictors (matrix or data.frame)}
+
+\item{Y}{target variable (vector whose length is equal to the number of rows
+in X)}
+
+\item{group.method}{group method. Default to \code{"adjclust"}}
+
+\item{summary.method}{summary method. Default to \code{"pls"}}
+
+\item{selection.method}{selection method. Default to \code{"none"} (no 
+selection performed)}
+
+\item{at}{number of groups targeted for output results (integer). Not used
+when \code{range.at} is not \code{NULL}}
+
+\item{range.at}{(vector of integer) sequence of the numbers of groups for
+output results}
+
+\item{seed}{random seed (integer)}
+
+\item{repeats}{number of repeats for the final random forest computation}
+
+\item{parallel}{not implemented yet}
+}
+\value{
+an object of class \code{"SFCB"} with elements: 
+  \item{dendro}{a dendrogram corresponding to the method chosen in
+  \code{group.method}}
+  \item{groups}{a list of length \code{length(range.at)} (or of length 1 if
+  \code{range.at == NULL}) that contains the clusterings of input variables
+  for the selected group numbers}
+  \item{summaries}{a list of the same length than \code{$groups} that 
+  contains the summarized predictors according to the method chosen in 
+  \code{summary.methods}}
+  \item{selected}{a list of the same length than \code{$groups} that contains
+  the names of the variable selected by \code{selection.method} if it is not
+  equal to \code{"none"}}
+  \item{mse}{a data.frame with \code{repeats} \eqn{\times}
+  \code{length($groups)} rows that contains Mean Squared Errors of the 
+  \code{repeats} random forests fitted for each number of groups}
+  \item{importance}{a list of the same length than \code{$groups} that 
+  contains a data.frame providing variable importances for the variables in
+  selected groups in \code{repeats} columns (one for each iteration of the
+  random forest method). When \code{summary.method == "basics"}, importance 
+  for mean and sd are provided in separated columns, in which case, the 
+  number of columns is equal to 2\code{repeats}}
+}
+\description{
+\code{SFCB} performs interval selection based on random forests
+}
+\author{
+{Remi Servien, \email{remi.servien@inrae.fr}\cr
+Nathalie Vialaneix, \email{nathalie.vialaneix@inrae.fr}}
+}

man/sparseRes.Rd

+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/sparseSIR.R
+\name{sparseRes}
+\alias{sparseRes}
+\alias{summary.sparseRes}
+\alias{print.sparseRes}
+\alias{sparseRes-class}
+\title{Print sparseRes object}
+\usage{
+\method{summary}{sparseRes}(object, ...)
+
+\method{print}{sparseRes}(x, ...)
+}
+\arguments{
+\item{object}{a \code{sparseRes} object}
+
+\item{...}{not used}
+
+\item{x}{a \code{sparseRes} object}
+}
+\description{
+Print a summary of the result of \code{\link{sparseSIR}} (
+\code{sparseRes} object)
+}
+\seealso{
+\code{\link{sparseSIR}}
+}
+\author{
+{Victor Picheny, \email{victor.picheny@inrae.fr}\cr
+Remi Servien, \email{remi.servien@inrae.fr}\cr
+Nathalie Vialaneix, \email{nathalie.vialaneix@inra.fr}}
+}

man/sparseSIR.Rd

+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/sparseSIR.R
+\name{sparseSIR}
+\alias{sparseSIR}
+\title{sparse SIR}
+\usage{
+sparseSIR(
+  object,
+  inter_len,
+  adaptive = FALSE,
+  sel_prop = 0.05,
+  parallel = FALSE,
+  ncores = NULL
+)
+}
+\arguments{
+\item{object}{an object of class \code{ridgeRes} as obtained from the 
+function \code{\link{ridgeSIR}}}
+
+\item{inter_len}{(numeric) vector with interval lengths}
+
+\item{adaptive}{should the function returns the list of strong zeros and non
+strong zeros (logical). Default to FALSE}
+
+\item{sel_prop}{used only when \code{adaptive = TRUE}. Fraction of the 
+coefficients that will be considered as strong zeros and strong non zeros.
+Default to 0.05}
+
+\item{parallel}{whether the computation should be performed in parallel or
+not. Logical. Default is FALSE}
+
+\item{ncores}{number of cores to use if \code{parallel = TRUE}. If left to 
+NULL, all available cores minus one are used}
+}
+\value{
+S3 object of class \code{sparseRes}: a list consisting of
+\itemize{
+   \item{\code{sEDR}}{ the estimated EDR space (a p x d matrix)}
+   \item{\code{alpha}}{ the estimated shrinkage coefficients (a vector having
+   a length similar to \code{inter_len})}
+   \item{\code{quality}}{ a vector with various qualities for the model (see
+   Details)}
+   \item{\code{adapt_res}}{ if \code{adaptive = TRUE}, a list of two vectors: 
+   \itemize{
+     \item{\code{nonzeros}}{ indexes of variables that are strong non zeros}
+     \item{\code{zeros}}{ indexes of variables that are strong zeros}
+   }}
+   \item{\code{parameters}}{ a list of hyper-parameters for the method: 
+   \itemize{
+     \item{\code{inter_len}}{ lengths of intervals}
+     \item{\code{sel_prop}}{ if \code{adaptive = TRUE}, fraction of the 
+     coefficients which are considered as strong zeros or strong non zeros}
+   }}
+   \item{\code{rSIR}}{ same as the input \code{object}}
+   \item{\code{fit}}{ a list for LASSO fit with:
+   \itemize{
+     \item{\code{glmnet}} result of the \code{\link[glmnet]{glmnet}} function
+     \item{\code{lambda}} value of the best Lasso parameter by CV
+     \item{\code{x}} exploratory variable values as passed to fit the model
+   }}
+ }
+ 
+ @details Different quality criteria used to select the best models among a
+ list of models with different interval definitions. Quality criteria are:
+ log-likelihood (\code{loglik}), cross-validation error as provided by the
+ function \code{\link[glmnet]{glmnet}}, two versions of the AIC (\code{AIC} 
+ and \code{AIC2}) and of the BIC (\code{BIC} and \code{BIC2}) in which the 
+ number of parameters is either the number of non null intervals or the 
+ number of non null parameters with respect to the original variables.
+}
+\description{
+\code{sparseSIR} performs the second step of the method (shrinkage of ridge
+SIR results
+}
+\examples{
+set.seed(1140)
+tsteps <- seq(0, 1, length = 200)
+nsim <- 100
+simulate_bm <- function() return(c(0, cumsum(rnorm(length(tsteps)-1, sd=1))))
+x <- t(replicate(nsim, simulate_bm()))
+beta <- cbind(sin(tsteps*3*pi/2), sin(tsteps*5*pi/2))
+beta[((tsteps < 0.2) | (tsteps > 0.5)), 1] <- 0
+beta[((tsteps < 0.6) | (tsteps > 0.75)), 2] <- 0
+y <- log(abs(x \%*\% beta[ ,1]) + 1) + sqrt(abs(x \%*\% beta[ ,2]))
+y <- y + rnorm(nsim, sd = 0.1)
+res_ridge <- ridgeSIR(x, y, H = 10, d = 2, mu2 = 10^8)
+res_sparse <- sparseSIR(res_ridge, rep(10, 20))
+
+}
+\references{
+{Picheny, V., Servien, R. and Villa-Vialaneix, N. (2016) 
+Interpretable sparse SIR for digitized functional data. 
+\emph{Statistics and Computing}, \strong{29}(2), 255--267.}
+}
+\seealso{
+\code{\link{ridgeSIR}}, \code{\link{project.sparseRes}}, 
+\code{\link{SISIR}}
+}
+\author{
+{Victor Picheny, \email{victor.picheny@inrae.fr}\cr
+Remi Servien, \email{remi.servien@inrae.fr}\cr
+Nathalie Vialaneix, \email{nathalie.vialaneix@inrae.fr}}
+}

man/tune.ridgeSIR.Rd

+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/ridgeSIR.R
+\name{tune.ridgeSIR}
+\alias{tune.ridgeSIR}
+\title{Cross-Validation for ridge SIR}
+\usage{
+tune.ridgeSIR(
+  x,
+  y,
+  listH,
+  list_mu2,
+  list_d,
+  nfolds = 10,
+  parallel = TRUE,
+  ncores = NULL
+)
+}
+\arguments{
+\item{x}{explanatory variables (numeric matrix or data frame)}
+
+\item{y}{target variable (numeric vector)}
+
+\item{listH}{list of the number of slices to be tested (numeric vector)}
+
+\item{list_mu2}{list of ridge regularization parameters to be tested 
+(numeric vector)}
+
+\item{list_d}{list of the dimensions to be tested (numeric vector)}
+
+\item{nfolds}{number of folds for the cross validation. Default is 10}
+
+\item{parallel}{whether the computation should be performed in parallel or
+not. Logical. Default is FALSE}
+
+\item{ncores}{number of cores to use if \code{parallel = TRUE}. If left to 
+NULL, all available cores minus one are used}
+}
+\value{
+a data frame with tested parameters and corresponding CV error and 
+estimation of R(d)
+}
+\description{
+\code{tune.ridgeSIR} performs a Cross Validation for ridge SIR estimation
+}
+\examples{
+set.seed(1115)
+tsteps <- seq(0, 1, length = 200)
+nsim <- 100
+simulate_bm <- function() return(c(0, cumsum(rnorm(length(tsteps)-1, sd=1))))
+x <- t(replicate(nsim, simulate_bm()))
+beta <- cbind(sin(tsteps*3*pi/2), sin(tsteps*5*pi/2))
+y <- log(abs(x \%*\% beta[ ,1])) + sqrt(abs(x \%*\% beta[ ,2]))
+y <- y + rnorm(nsim, sd = 0.1)
+list_mu2 <- 10^(0:10)
+listH <- c(5, 10)
+list_d <- 1:4
+set.seed(1129)
+\dontrun{
+res_tune <- tune.ridgeSIR(x, y, listH, list_mu2, list_d, 
+                          nfolds = 10, parallel = TRUE)}
+
+}
+\references{
+{Picheny, V., Servien, R. and Villa-Vialaneix, N. (2016) 
+Interpretable sparse SIR for digitized functional data.
+\emph{Statistics and Computing}, \strong{29}(2), 255--267.}
+}
+\seealso{
+\code{\link{ridgeSIR}}
+}
+\author{
+{Victor Picheny, \email{victor.picheny@inrae.fr}\cr
+Remi Servien, \email{remi.servien@inrae.fr}\cr
+Nathalie Vialaneix, \email{nathalie.vialaneix@inrae.fr}}
+}