dcsvm (version 0.0.1)

cv.dcsvm: Cross-Validation for Sparse Density-Convoluted SVM

Description

Performs cross-validation for the sparse density-convoluted SVM to estimate the optimal tuning parameter lambda.

Usage

cv.dcsvm(x, y, lambda = NULL, hval = 1, 
  pred.loss = c("misclass", "loss"), nfolds = 5, foldid, ...)

Value

A cv.dcsvm object is returned, which includes the cross-validation fit:

lambda

The lambda sequence used in dcsvm.

cvm

A vector of length length(lambda) for the mean cross-validated error.

cvsd

A vector of length length(lambda) for estimates of standard error of cvm.

cvupper

The upper curve: cvm + cvsd.

cvlower

The lower curve: cvm - cvsd.

nzero

Number of non-zero coefficients at each lambda.

name

"Mis-classification error", for plotting purposes.

dcsvm.fit

A fitted dcsvm object using the full data.

lambda.min

The lambda incurring the minimum cross-validation error cvm.

lambda.1se

The largest value of lambda such that error is within one standard error of the minimum.

cv.min

The minimum cross-validation error.

cv.1se

The cross-validation error associated with lambda.1se.

Arguments

x

A matrix of predictors, i.e., the x matrix used in dcsvm.

y

A vector of binary class labels, i.e., the y used in dcsvm.

lambda

Default is NULL, and the sequence generated by dcsvm is used. User can also provide a new lambda sequence for cross-validation.

hval

The bandwidth parameter for kernel smoothing. Default is 1.

pred.loss

"misclass" for classification error, "loss" for the density-convoluted SVM loss.

nfolds

The number of folds. Default is 5. The allowable range is from 3 to the sample size. Larger nfolds increases computational time.

foldid

An optional vector with values between 1 and nfold, representing the fold indices for each observation. If supplied, nfolds can be missing.

...

Other arguments that can be passed to dcsvm.

Details

Cross-Validation for Sparse Density-Convoluted SVM

Conducts a k-fold cross-validation for dcsvm and returns the suggested values of the L1 parameter lambda.

This function runs dcsvm on the sparse density-convoluted SVM by excluding each fold in turn, then computes the mean cross-validation error and standard deviation. It is adapted from the cv functions in the gcdnet and glmnet packages.

See Also

dcsvm, plot.cv.dcsvm, predict.cv.dcsvm, and coef.cv.dcsvm methods.

Examples

data(colon)
colon$x <- colon$x[ ,1:100] # Use only the first 100 columns for this example
n <- nrow(colon$x)
set.seed(1)
id <- sample(n, trunc(n / 3))
cvfit <- cv.dcsvm(colon$x[-id, ], colon$y[-id], lam2=1, nfolds=5)
plot(cvfit)
predict(cvfit, newx=colon$x[id, ], s="lambda.min")