fasta: Fast Adaptive Shrinkage/Thresholding Algorithm

A collection of acceleration schemes for proximal gradient methods for estimating penalized regression parameters described in Goldstein, Studer, and Baraniuk (2016) <arXiv:1411.3406>. Schemes such as Fast Iterative Shrinkage and Thresholding Algorithm (FISTA) by Beck and Teboulle (2009) <doi:10.1137/080716542> and the adaptive stepsize rule introduced in Wright, Nowak, and Figueiredo (2009) <doi:10.1109/TSP.2009.2016892> are included. You provide the objective function and proximal mappings, and it takes care of the issues like stepsize selection, acceleration, and stopping conditions for you.

Getting started

Package details

AuthorEric C. Chi [aut, cre, trl, cph], Tom Goldstein [aut] (MATLAB original, https://www.cs.umd.edu/~tomg/projects/fasta/), Christoph Studer [aut], Richard G. Baraniuk [aut]
MaintainerEric C. Chi <[email protected]>
LicenseMIT + file LICENSE
Version0.1.0
Package repositoryView on CRAN
Installation Install the latest version of this package by entering the following in R:
install.packages("fasta")

Try the fasta package in your browser

Any scripts or data that you put into this service are public.

fasta documentation built on April 10, 2018, 5:05 p.m.