cv.rcar: Cross-validation for rcar
In azemi/RCAR: Lasso Regularized CLASS-ASSOCIATION RULES BASED CLASSIFIER

Description Usage Arguments Details Value

View source: R/cv_rcar.R

Does 10-fold cross-validation for RCAR over an automatically choosen sequence of values of lambda.

1	cv.rcar(data, y.rank = 1, s = 0.3, c = 0.7, maxl = 6)

`data:`	A categorical data set with binary response variable; each row is an observation vector.
`y.rank:`	The rank of the response variable.
`s:`	A user supplied value for the minimal support of an item set.
`c:`	A user supplied value for the minimal confidence of class-association rules.
`maxl:`	A user specified value for the maximal number of items per item set (default: 6 items).

10-folds cross validation is calculated over an automatically determined sequence of lambda values ranging from the value of lambda such that all the coefficients are zero down to value of lambda whose the deviance do not change from lambda to the next. Note also that the results of cv.rcar are random, since the folds are selected randomly. Users can reduce this randomness by running cv.rcar many times, and averaging the error curves.

An object of class "cv.rcar" is returned, which is a list with the ingredients of the cross-validation fit plus the elements of the most accurate model.

azemi/RCAR documentation built on May 7, 2019, 2:54 a.m.