Beyond L1: Faster and Better Sparse Models with skglm
Machine Learning
2023-03-09 v2 Machine Learning
Abstract
We propose a new fast algorithm to estimate any sparse generalized linear model with convex or non-convex separable penalties. Our algorithm is able to solve problems with millions of samples and features in seconds, by relying on coordinate descent, working sets and Anderson acceleration. It handles previously unaddressed models, and is extensively shown to improve state-of-art algorithms. We provide a flexible, scikit-learn compatible package, which easily handles customized datafits and penalties.
Cite
@article{arxiv.2204.07826,
title = {Beyond L1: Faster and Better Sparse Models with skglm},
author = {Quentin Bertrand and Quentin Klopfenstein and Pierre-Antoine Bannier and Gauthier Gidel and Mathurin Massias},
journal= {arXiv preprint arXiv:2204.07826},
year = {2023}
}