Cal_grpWTs: An auxilary function calculating the group weighting matrix...

In MSGLasso: Multivariate Sparse Group Lasso for the Multivariate Multiple Linear Regression with an Arbitrary Group Structure

Description

An auxilary function calculating the group weighting matrix grpWTs required when calling the MSGlasso function.

Usage

Cal_grpWTs(P, Q, G, R, gmax, PQ.grps)

Arguments

P	a positive interger indicating number of predictor variables
Q	a positive interger indicating number of response variables
G	a positive interger indicating number of predictor groups
R	a positive interger indicating number of response groups
gmax	a positive interger indicating the max number of different groups a single variable (either a predictor or response variable) belongs to.
PQ.grps	a matrix of (p+q) by (gmax+1), with each row starting with group indicators that row variable belongs to, and followed by 999's till the row is filled.

Details

Generates the required input group weighting matrix grpWTs when calling the main MSGlasso function. The grpWTs is a g by r matrix containing the adaptive weighting scores for each group. MSGLasso.grpWTs use the square root of the group size (number of entries the group contains) as the weight for each group.

Value

A list with one components:

grpWTs

the grpWTs matrix generated

Author(s)

Yanming Li, Bin Nan, Ji Zhu

References

Y. Li, B. Nan and J. Zhu (2015) Multivariate sparse group lasso for the multivariate multiple linear regression with an arbitrary group structure. Biometrics. DOI: 10.1111/biom.12292

Examples

###########################################################################
## generating the grp.WTs matrix for an overlapping group structure
###########################################################################

P <- 200
Q <- 200
G <- 10
R <- 10
gmax <- 1

GarrStarts <-c(0,20,40,60,80,100,120,140,160,180)
GarrEnds <-c(19,39,59,79,99,119,139,159,179,199)
RarrStarts <-c(0,20,40,60,80,100,120,140,160,180)
RarrEnds <-c(19,39,59,79,99,119,139,159,179,199)

tmp <- FindingPQGrps(P, Q, G, R, gmax, GarrStarts, GarrEnds, RarrStarts, RarrEnds)
PQ.grps <- tmp$PQgrps

tmp1 <- Cal_grpWTs(P, Q, G, R, gmax, PQ.grps)
grp.WTs <- tmp1$grpWTs

MSGLasso documentation built on May 1, 2019, 6:51 p.m.