extractFeatures: Extracts features which have been choosen by the...

In pathClass: Classification using biological pathways as prior knowledge

Description

This function extracts the features which have been selected by the classifiers during the cross-validation along with the number of times they have been choosen. When, for example, performing a 5 times repeated 10-fold cross-validation the maximum number a feature can be choosen is 50.

Usage

  extractFeatures(res, toFile = FALSE,
    fName = "ClassificationFeatures.csv")

Arguments

res	A result of crossval.
toFile	Should the results be printed into a CSV-file.
fName	the name of the file to save the results in.

Value

a data.frame indicating the number of times a feature has been choosen.

Author(s)

Marc Johannes [e-mail]

Examples

## Not run: 
library(Biobase)
data(sample.ExpressionSet)
x <- t(exprs(sample.ExpressionSet))
y <- factor(pData(sample.ExpressionSet)$sex)
res.rfe <- crossval(x,y,DEBUG=TRUE,theta.fit=fit.rfe,folds=2,repeats=1,parallel=TRUE,Cs=10^(-3:3))
extractFeatures(res.rfe, toFile=FALSE)

## End(Not run)

pathClass documentation built on May 29, 2017, 11:44 p.m.