plsrf_x_pv: Classification based on pre-validated PLS dimension reduction...
In MAclinical: Class prediction based on microarray data and clinical parameters

Description Usage Arguments Details Value Author(s) References See Also Examples

View source: R/plsrf_x_pv.r

This function builds a prediction rule based on the learning data (microarray predictors only) and applies it to the test data. The classifier consists of two steps: PLS dimension reduction with pre-validation step for summarizing microarray data, and random forests applied to the obtained PLS components. See Boulesteix et al (2008) for more details.

The function plsrf_x_pv uses the functions cforest and varimp from the package party and the function pls.regression from the package plsgenomics.

1 2	plsrf_x_pv(Xlearn,Zlearn=NULL,Ylearn,Xtest,Ztest=NULL,ncomp=0:3, ordered=NULL,nbgene=NULL,fold=10,...)

`Xlearn`	A nlearn x p matrix giving the microarray predictors for the learning data set.
`Zlearn`	A nlearn x q matrix giving the clinical predictors for the learning data set. This argument is ignored.
`Ylearn`	A numeric vector of length nlearn giving the class membership of the learning observations, coded as 0,...,K-1 (where K is the number of classes).
`Xtest`	A ntest x p matrix giving the microarray predictors for the test data set.
`Ztest`	A ntest x q matrix giving the clinical predictors for the test data set. This argument is ignored.
`ncomp`	A numeric vector giving the candidate numbers of pre-validated PLS components. All numbers must be >0.
`ordered`	A vector of length p giving the order of the microarray predictors in terms of relevance for prediction. For instance, if the three first elements of `ordered` are 30,2,2400, it means that the most relevant genes are the genes in the 30th, 2nd and 2400th columns of the gene expression data matrix `Xlearn`. Note: if `ordered=NULL`, the columns of `Xlearn` and `Xtest` are assumed to be already ordered.
`nbgene`	The number of genes to be selected for use in dimension reduction. Default is `nbgene=NULL`, in which case all genes are used.
`fold`	The number of folds for the pre-validation step. See Boulesteix et al (2008) for more details. The default is `fold=10`.
`...`	Other arguments to be passed to the function `cforest_control` from the `party` package.

See Boulesteix et al (2008).

A list with the elements:

`prediction`	A numeric vector of length `nrow(Xtest)` giving the predicted class for each observation from the test data set.
`importance`	The variable importance information output by the function `varimp` from the package `party` for the corresponding forest.
`bestncomp`	The best number of pre-validated PLS components, as obtained using the model selection method based on the out-of-bag error.
`OOB`	A numeric vector of length `ncomp` giving the out-of-bag error of the forest constructed with the corresponding number of pre-validated PLS components.

Anne-Laure Boulesteix (http://www.ibe.med.uni-muenchen.de/organisation/mitarbeiter/020_professuren/boulesteix/eng.html)

Boulesteix AL, Porzelius C, Daumer M, 2008. Microarray-based classification and clinical predictors: On combined classifiers and additional predictive value. Bioinformatics 24:1698-1706.

Tibshirani R, Efron B, 2002. Pre-validation and inference in microarrays. Stat. Appl. Genet. Mol. Biol. 1:1.

testclass, testclass_simul, simulate, plsrf_x, plsrf_xz, plsrf_xz_pv, rf_z, logistic_z, svm_x.

# load MAclinical library
# library(MAclinical)

# Generating xlearn, zlearn, ylearn, xtest, ztest
xlearn<-matrix(rnorm(3000),30,100)
ylearn<-sample(0:1,30,replace=TRUE)
xtest<-matrix(rnorm(2000),20,100)

my.prediction1<-plsrf_x_pv(Xlearn=xlearn,Ylearn=ylearn,Xtest=xtest)

ordered<-sample(100)
my.prediction2<-plsrf_x(Xlearn=xlearn,Ylearn=ylearn,Xtest=xtest,ordered=ordered,nbgene=20)
my.prediction3<-plsrf_x_pv(Xlearn=xlearn,Ylearn=ylearn,Xtest=xtest,fold=30)