screening.group: Screening.group

In jbecu/ridgeAdap: Adaptive Ridge for variable selection

Description

Screening.group

Usage

screening.group(X,y,lambda = NULL,alpha=0,nlam=20,nfolds=10,group=1:ncol(X),
thresh=0.001,scale=FALSE,center=TRUE)

Arguments

X	a matrix object with n*p size representing the data
y	a vector of size p representing the output variable
lambda	a vector representing the l1 penalty. Default is NULL and automatically defined by SGL.
alpha	a vector representing the α parameter for the sparse group lasso. If alpha = 1 that correspond to the lasso solution, if alpha = 0 that correspond to the group lasso and otherwise to the sparse group lasso solution. Default is 0.
nlam	an integer representing the size of the lambda grid if lambda is NULL. Default is 20.
nfolds	an integer representing the number of folds for the k-folds cross-validation used to choose l1 and l2 penatly. Default is 10 and must be greater than 2.
group	a vector of size p representing the group index for each variables. Default is 1:ncol(X) which represent the particular case whit no group. Warning this vector must be in ascending group order and variables in X needs to be ordered in this sense.
thresh	convergence threshold for change in beta.
scale	a boolean to indicate if X and y will be scaled. Default is TRUE.
center	a boolean to indicate if X and y will be centered. Default is TRUE.

Value

A object of class 'screening' with 5 elements

beta.min	a vector of size p with the elastic.net estimates
S	a vector contaning the index of variable which $B_j != 0$
lambda1.min	l1 penalty choose by CV
alpha.min	alpha mixing parameter choose by CV
ref	a list with two component contains respectively the X and y variance and mean used to scaling and/or centering the data.

Note

screening.group requires SGL package.

Author(s)

JM BECU

Examples

 # see ridgeAdap help

jbecu/ridgeAdap documentation built on May 18, 2019, 5:58 p.m.