sigCVError: Mean external cross validation (k-fold) error calculation.

In sigFeature: sigFeature: Significant feature selection using SVM-RFE & t-statistic

Description

The function "sigCVError()" computes the mean external cross validation (k-fold) errors and the standered deviation of the errors.

Usage

sigCVError(i, results, input)

Arguments

i	number of features/genes from top of the ranked list.
results	the results is produced from sigFeatureRanking.enfold() function.
input	vector of class labels -1 or 1's (for n samples/patients). n-by-d data matrix to train (n chips/patients, d features/genes).

Details

The features which are produced on the basis of frequency are used here to enumerate mean external cross validation (k-fold) errors and the standard deviation of the errors. Training and test samples are same which are initially produced after splitting the main sample set in to k-fold. In each iteration k-1 folds are considered as training samples and remaining one fold is considered as test samples.In this external cross validation procedure, feature numbers are increased one by one by using the expression values from training dataset as well as test dataset. After that, traning samples are trained to test the tesing samples dynamically. The number of un classified samples are averaged and are called as external cross validation error rate.

Value

error

Return the error list.

Author(s)

Pijush Das <[e-mail]>, et al.

References

Zhang, H. H., Ahn, J., Lin, X. and Park, C. (2006). Gene selection using support vector machines with nonconvex penalty. Bioinformatics, 22, pp. 88-95.

See Also

findgacv.scad, predict.penSVM, sim.data

Examples

#Example for sigCVError()
#Data set taken from GSE2280
#library(SummarizedExperiment)
#data(ExampleRawData, package="sigFeature")

#x  <- t(assays(ExampleRawData)$counts)
#y  <- colData(ExampleRawData)$sampleLabels

#inputdata <- data.frame(y=as.factor(y) ,x=x)

#For 10 fold cross validation.
#results = sigFeature.enfold(x, y, "kfold",10)

#Find out the frequency of the top 400 features selected by every iteration.
#FeatureBasedonFrequency <- sigFeatureFrequency(results, 400, 400, pf=FALSE)
#str(FeatureBasedonFrequency[1])


#To run the code given bellow will take huge time.
#featsweepSigFe = lapply(1:400, sigCVError, FeatureBasedonFrequency, inputdata)

#Thus the process data is given below.
data("featsweepSigFe")
str(featsweepSigFe[1])

sigFeature documentation built on Nov. 8, 2020, 8:28 p.m.