netClass (version 1.0)

stsvm.cv: Cross validation for smoothed t-statistic to select significant top ranked differential expressed genes

Description

Cross validation for smoothed t-statistic to select significant top ranked differential expressed genes

Usage

stsvm.cv(x=x, x.mi=NULL,y=y, folds=5,Gsub=matrix(1,100,100),repeats=3, parallel=FALSE,
	cores=2, DEBUG=TRUE, pt.pvalue = 0.05, op = 0.85, pt.step = 1000, a = 1, p=2,
	allF=TRUE, seed=1234, Cs = 10^c(-3:3) )

Arguments

x

A p x n matrix of expression measurements with p samples and n genes.

x.mi

A p x m matrix of expression measurements with p samples and m miRNAs.

y

A factor of length p comprising the class labels.

folds

Folds number of folds to perform

Gsub

An adjacency matrix that represents the underlying biological network.

repeats

Number of how often to repeat the x-fold cross-validation

parallel

Use parallel computing or not

cores

Number of cores will used when parallel is TRUE

DEBUG

Show debugging information in screen more or less.

pt.pvalue

Cut off p-value of permutation test

op

Pptimal on top op

pt.step

permutation test steps

a

constant value of random walk kernel

p

random walk step(s) of random walk kernel

allF

Using all features (TRUE) or only these genes mapped to prior information (FALSE).

seed

seed for random sampling.

Cs

Soft-margin tuning parameter of the SVM. Defaults to 10^c(-3:3).

Value

• a LIST for Cross-Validation results
• auc
  The AUC values of each test fold
• fits
  The tranined models for traning folds
• feat
  The feature selected by each by the fits
• labels
  the original lables for training

References

Yupeng Cun, Holger Frohlich (2013) Network and Data Integration for Biomarker Signature Disvovery via Network Smoothed T-Statistics

Examples

library(netClass)
data(expr)
data(ad.matrix)
x <- expr$genes
y <- expr$y

r.stsvm <- stsvm.cv(x=x[,1:100],x.mi=NULL,y=y,folds=3,Gsub=ad.matrix,repeats=1, parallel=FALSE,
			 cores=2,DEBUG=TRUE,pt.pvalue=0.05,op=0.9,pt.step=3,a=1,p=2,allF=TRUE,
			 seed=1234,Cs=10^(-3:3))