smart (version 1.0.1)

gnsc.cv: A function to cross-validate the Group Nearest Shrunken Centroid Classifier

Description

A function to cross-validate the Group Nearest Shrunken Centroid Classifier produced by gnsc.train

Usage

gnsc.cv(fit, x, y = NULL, z = NULL, nfold = NULL, folds = NULL, verbose = T)

Arguments

fit

The result of a call to gnsc.train

x

The test data matrix (variables in the rows, samples in the columns).

y

The test class labels for samples, must have the same length as the column length of x.

z

The test class labels for variables, must have the same length as the row length of x.

nfold

Number of cross-validation folds. The default value is the smallest class size.

folds

The fold lables for each sample, must have the same length as y and max(folds)=nfold. The default value is sample(1:nfold,n,replace=T), here n is the sample size.

verbose

If verbose = FALSE, tracing information printing is disabled. The default value is TRUE.

Value

An object with S3 class "gnsccv" is returned:

lambda

A vector of the thresholds tried in the shrinkage

nlambda

The number of thresholds tried in the shrinkage

lambda.min

The index of the threshold which achieves the lowest cross-validation error

errors

The number of cross-validation errors for each threshold value

nonzero

The number of variables that survived the thresholding

Thresh.mat

A list of estimated tilde{mu}_{mk}. See Yang, et.al (2012) for details

Details

gnsc.cv carries out a cross-validation for Group Nearest Shrunken Centroid Classifier.

References

1.Juemin Yang, Fang Han, Rafa Irizarry, and Han Liu. Gene Context Analysis on Large-scale Genomic Data. Technical Report, Johns Hopkins University, 2012 2.Robert Tibshirani, Trevor Hastie, Balasubramanian Narasimhan, and Gilbert Chu. Diagnosis of multiple cancer types by shrunken centroids of gene expression PNAS, 99: 6567-6572.

See Also

gnsc.train

Examples

set.seed(120)
x <- matrix(rnorm(1000*20),ncol=20)
y <- sample(c(1:4),size=20,replace=TRUE)
z <- sample(c(1:10),size=1000,replace=TRUE)
fit=gnsc.train(x, col.struc=y, row.struc=z,lambda.max=5, nlambda=20)
fit
plot(fit)
fit.cv=gnsc.cv(fit,x,y,z)
fit.cv
plot(fit.cv)