nominalCV: Cross-validation of a nominal model with penalty

In micbtz/regIRT: Regularized estimation of IRT models

Description

Selection of the tuning parameter of a nominal model with penalty using cross-validation.

Usage

nominalCV(object, K, trace = FALSE)

Arguments

object	output of function nominalmod.
K	Number of folds.
trace	Trace information.

Value

A list with components:

est	list containing the parameter estimates on the training sets. Each component corresponds to a different value of lambda, and it is a matrix with K (= number of folds) rows and J (= number of parameters) columns.
lik	list containing the negative log-likelihood (not penalized) computed on the validation set. Each component corresponds to a different value of lambda, and it is a vector with length K.
lambda	vector of tuning parameters.
sel	integer value indicating the selection.
lambdasel	value of lambda selected.
par	matrix of parameter estimates obtained with function nominalmod. Columns correspond to different values of lambda.
data	dataset.
D	number of dimensions.

See Also

nominalmod, regIRT, regPath

micbtz/regIRT documentation built on July 6, 2019, 2:37 p.m.