coef.grpSLOPE: Extract model coefficients

In agisga/grpSLOPE: Group Sorted L1 Penalized Estimation

Extract model coefficients

Description

Extract the regression coefficients from a grpSLOPE object, either on the scale of the normalized design matrix (i.e., columns centered and scaled to unit norm), or on the original scale.

Usage

## S3 method for class 'grpSLOPE'
coef(object, scaled = TRUE, ...)

Arguments

object	A grpSLOPE object
scaled	Should the coefficients be returned for the normalized version of the design matrix?
...	Potentially further arguments passed to and from methods

Details

If scaled is set to TRUE, then the coefficients are returned for the normalized version of the design matrix, which is the scale on which they were computed. If scaled is set to FALSE, then the coefficients are transformed to correspond to the original (unaltered) design matrix. In case that scaled = FALSE, an estimate for the intercept term is returned with the other coefficients. In case that scaled = TRUE, the estimate of the intercept is always equal to zero, and is not explicitly provided.

Value

A named vector of regression coefficients where the names signify the group that each entry belongs to

Examples

set.seed(1)
A   <- matrix(rnorm(100^2), 100, 100)
grp <- rep(rep(letters[1:20]), each=5)
b   <- c(rep(1, 20), rep(0, 80))
y   <- A %*% b + rnorm(10) 
result <- grpSLOPE(X=A, y=y, group=grp, fdr=0.1)
head(coef(result), 8)
#       a_1       a_2       a_3       a_4       a_5       b_1       b_2       b_3 
#  7.942177  7.979269  8.667013  8.514861 10.026664  8.963364 10.037355 10.448692 
head(coef(result, scaled = FALSE), 8)
# (Intercept)         a_1         a_2         a_3         a_4         a_5         b_1         b_2 
#  -0.4418113   0.8886878   0.8372108   0.8422089   0.8629597   0.8615827   0.9323849   0.9333445 

agisga/grpSLOPE documentation built on June 3, 2023, 10:16 a.m.