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

This package contains two functions. One is the thresholding function for the scalar penalized least squares problem subject to a power penalty. The second is a function that solves a penalized regression problem subject to a power penalty via coordinate descent. For powers greater than zero and less than or equal to 1, the coordinate descent function implements the algorithm given by Marjanovic and Solo (2014). As a special case, the coordinate descent function solves the Lasso regression problem.

Getting started

Package details

AuthorMaryclare Griffin
MaintainerMaryclare Griffin <[email protected]>
Package repositoryView on GitHub
Installation Install the latest version of this package by entering the following in R:
maryclare/powopt documentation built on July 27, 2017, 1:05 a.m.