cv.gren: Performance cross-validation of group-regularized logistic...
In gren: Adaptive Group-Regularized Logistic Elastic Net Regression

Description Usage Arguments Details Value Author(s) References See Also Examples

Function that cross-validations for performance estimation of gren models.

cv.gren(x, y, m=rep(1, nrow(x)), unpenalized=NULL, partitions=NULL, alpha=0.5, 
        lambda=NULL, intercept=TRUE, monotone=NULL, psel=TRUE, compare=TRUE, 
        posterior=FALSE, nfolds=nrow(x), foldid=NULL, trace=TRUE,
        control=list(epsilon=0.001, maxit=500, maxit.opt=1000, maxit.vb=100),
        keep.pred=TRUE, fix.lambda=FALSE, nfolds.out=nrow(x), foldid.out=NULL,
        type.measure=c("auc", "deviance", "class.error"))

`x`	See `gren`.
`y`	See `gren`.
`m`	See `gren`.
`unpenalized`	See `gren`.
`partitions`	See `gren`.
`alpha`	See `gren`.
`lambda`	See `gren`.
`intercept`	See `gren`.
`monotone`	See `gren`.
`psel`	See `gren`.
`compare`	See `gren`.
`posterior`	See `gren`.
`nfolds`	See `gren`.
`foldid`	See `gren`.
`trace`	if `TRUE`, progress of the cross-validation is printed.
`control`	See `gren`.
`keep.pred`	`logical`. If `TRUE` the cross-validated predictions are saved.
`fix.lambda`	`logical`. If `TRUE` `lambda` is cross-validated only once and used in every fold.
`nfolds.out`	`numeric` that gives the number of folds to use in the performance cross-validation. Default is `nrow(x)`.
`foldid.out`	optional `numeric` vector of length `nrow(x)` with the performance cross-validation fold assignments of the observations.
`type.measure`	`numeric` that gives the performance measures to calculate. Currently possible are AUC, deviance, and misclassification error.

cv.gren is a convenience function that gives cross-validated predictions. Performance measures are optionally calculated with these predictions. cv.gren is more efficient than simply looping over the folds, since it uses the final estimates of the previous fold as starting values for the next fold. This substantially reduces computation time.

Function returns a list of length two with the following components:

`groupreg`	`list` with the cross-validated predictions of length `nrow(x)` and possibly performance metrics of the group-regularized model.
`regular`	`list` with the cross-validated predictions of length `nrow(x)` and possibly performance metrics of the regular model.

Magnus M. Münch <m.munch@vumc.nl>

Münch, M.M., Peeters, C.F.W., van der Vaart, A.W., and van de Wiel, M.A. (2018). Adaptive group-regularized logistic elastic net regression. arXiv:1805.00389v1 [stat.ME].

gren, predict.gren, coef.gren

## Create data
p <- 1000
n <- 100
set.seed(2018)
x <- matrix(rnorm(n*p), ncol=p, nrow=n)
beta <- c(rnorm(p/2, 0, 0.1), rnorm(p/2, 0, 1))
m <- rep(1, n)
y <- rbinom(n, m, as.numeric(1/(1 + exp(-x %*% as.matrix(beta)))))
partitions <- list(groups=rep(c(1, 2), each=p/2))

## calculate cross-validated predictions and performance measures
fit.cv.gren <- cv.gren(x, y, m, partitions=partitions, fix.lambda=TRUE)