WilcoxCV: an R package for fast variable selection in cross-validation

Abstract

Our method is implemented in the freely available R package WilcoxCV which can be downloaded from the Comprehensive R Archive Network at http://cran.r-project.org/src/contrib/Descriptions/WilcoxCV.html.

Evaluating Microarray-based Classifiers: An Overview

“…The same holds for PCA dimension reduction (R function prcomp). Here is a list of specific R packages that are of particular interest for microarray-based classification and freely available without registration.pamr package for PAM (Tibshirani et al 2002)penalized package for penalized regression approaches: LASSO, L 2 (Goeman, 2007)glmpath package for LASSO regression (Park and Hastie, 2007)rda package for regularized discriminant analysis (Guo et al 2007)plsgenomics package for PLS-based classification (Boulesteix, 2004; Fort and Lambert-Lacroix, 2005)gpls package for generalized partial least squares classification (Ding and Gentleman, 2005)e1071 package for SVMrandomForest for random forests classification (Diaz-Uriarte and de Andrés, 2006)logitBoost package for logitBoost (Dettling and Bühlmann, 2003)BagBoosting package for bagboosting (Dettling, 2004)MADE4 package for classification by the “between-group analysis” (BGA) dimension reduction method (Culhane et al 2005)pdmclass package for classification using penalized discriminant methods (Ghosh, 2003)MLInterfaces package including unifying functions for cross-validation and validation on test data in combination with various classifiersMCRestimate package for fair comparison and evaluation of classification methods (Ruschhaupt et al 2004)Packages including functions for gene selection aregenefilter package including a function that computes t-tests quicklyWilcoxCV package for fast Wilcoxon based variable selection in cross-validation (Boulesteix, 2007)varSelRF R package for variable selections with random forests (Diaz-Uriarte and de Andrés, 2006)GALGO R package for variable selection with genetic algorithms (Trevino and Falciani, 2006) (http://www.bip.bham.ac.uk/vivo/galgo/AppNotesPaper.htm). …”

Repeated double cross validation

“…Widely used for this purpose are traditional versions of not only cross validation (CV) [4], but also Monte Carlo CV [5][6][7], and moving window CV [8], as well as bootstrap techniques [9,10]. It is important, however, to note that CV-based methods are not always optimal, particularly when dealing with data obtained by experimental design [11].…”

An Alternative Dimension Reduction Approach to Supervised Principal Components Analysis in High Dimensional Survival Data

“…The greatest value for the logarithmic likelihood ratio statistics is selected as the threshold value. 6,7,23 SUPERVISED PRINCIPAL COMPONENT ANALYSIS Principal components in SPCA, proposed by Bair and Tibshirani (2004) for gene expression data analysis, are estimated from a subset of genes which are related to the dependent variables. 6 Let X which is formed of p variables for n observations be an n x p matrix of independent variables and y be the n vector of dependent variable.…”