powObj: Penalized likelihood objective function

In maryclare/powopt: Solving Penalized Regression Problems with Power Penalty by Coordinate Descent

Penalized likelihood objective function

Description

Letting Q =X'X and l = X'y, gives the value:

-((β'Qβ - 2β'l)/(2σ^2) + λ ∑_{j = 1}^p |β_j|^q)

for fixed z, λ > 0 and q > 0.

Usage

powObj(beta, Q, l, sigmasq = 1, lambda = 1, q = 1)

Arguments

beta	length p vector of coefficient values to compute objective function for
Q	p\timesp matrix corresponding to X'X
l	length p vector corresponding to X'y
sigmasq	scalar noise variance
lambda	scalar multiplier applied to penalty
q	scalar power of penalty function

maryclare/powopt documentation built on Feb. 5, 2023, 6:05 p.m.