R: Cross-validation for grpreg

cv.grpreg {grpreg}

R Documentation

Cross-validation for grpreg

Description

Performs k-fold cross validation for penalized regression models with grouped covariates over a grid of values for the regularization parameter lambda.

Usage

cv.grpreg(X, y, group=1:ncol(X), ..., nfolds=10, seed, cv.ind,
returnY=FALSE, trace=FALSE)

Arguments

`X`	The design matrix, as in `grpreg`.
`y`	The response vector (or matrix), as in `grpreg`.
`group`	The grouping vector, as in `grpreg`.
`...`	Additional arguments to `grpreg`.
`nfolds`	The number of cross-validation folds. Default is 10.
`seed`	You may set the seed of the random number generator in order to obtain reproducible results.
`cv.ind`	Which fold each observation belongs to. By default the observations are randomly assigned by `cv.grpreg`.
`returnY`	Should `cv.grpreg` return the fitted values from the cross-validation folds? Default is FALSE; if TRUE, this will return a matrix in which the element for row i, column j is the fitted value for observation i from the fold in which observation i was excluded from the fit, at the jth value of lambda.
`trace`	If set to TRUE, cv.grpreg will inform the user of its progress by announcing the beginning of each CV fold. Default is FALSE.

Details

The function calls grpreg nfolds times, each time leaving out 1/nfolds of the data. The cross-validation error is based on the residual sum of squares when family="gaussian" and the deviance when family="binomial" or family="poisson".

For Gaussian and Poisson responses, the folds are chosen according to simple random sampling. For binomial responses, the numbers for each outcome class are balanced across the folds; i.e., the number of outcomes in which y is equal to 1 is the same for each fold, or possibly off by 1 if the numbers do not divide evenly.

As in grpreg, seemingly unrelated regressions/multitask learning can be carried out by setting y to be a matrix, in which case groups are set up automatically (see grpreg for details), and cross-validation is carried out with respect to rows of y. As mentioned in the details there, it is recommended to standardize the responses prior to fitting.

Value

An object with S3 class "cv.grpreg" containing:

`cve`	The error for each value of `lambda`, averaged across the cross-validation folds.
`cvse`	The estimated standard error associated with each value of for `cve`.
`lambda`	The sequence of regularization parameter values along which the cross-validation error was calculated.
`fit`	The fitted `grpreg` object for the whole data.
`min`	The index of `lambda` corresponding to `lambda.min`.
`lambda.min`	The value of `lambda` with the minimum cross-validation error.
`null.dev`	The deviance for the intercept-only model.
`pe`	If `family="binomial"`, the cross-validation prediction error for each value of `lambda`.

Author(s)

Patrick Breheny <patrick-breheny@uiowa.edu>

Examples

data(Birthwt)
X <- Birthwt$X
y <- Birthwt$bwt
group <- Birthwt$group

cvfit <- cv.grpreg(X, y, group)
plot(cvfit)
summary(cvfit)
coef(cvfit) ## Beta at minimum CVE

cvfit <- cv.grpreg(X, y, group, penalty="gel")
plot(cvfit)
summary(cvfit)

[Package grpreg version 3.1-4 Index]