Saved in:
Bibliographic Details
Main Authors: Schwendinger, Benjamin, Schwendinger, Florian, Vana-Gür, Laura
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2404.16560
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917147797618688
author Schwendinger, Benjamin
Schwendinger, Florian
Vana-Gür, Laura
author_facet Schwendinger, Benjamin
Schwendinger, Florian
Vana-Gür, Laura
contents In this paper, we show how mixed-integer conic optimization can be used to combine feature subset selection with holistic generalized linear models to fully automate the model selection process. Concretely, we directly optimize for the Akaike and Bayesian information criteria while imposing constraints designed to deal with multicollinearity in the feature selection task. Specifically, we propose a novel pairwise correlation constraint that combines the sign coherence constraint with ideas from classical statistical models like Ridge regression and the OSCAR model.
format Preprint
id arxiv_https___arxiv_org_abs_2404_16560
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Automated Model Selection for Generalized Linear Models
Schwendinger, Benjamin
Schwendinger, Florian
Vana-Gür, Laura
Machine Learning
Optimization and Control
68T05
G.3; G.4
In this paper, we show how mixed-integer conic optimization can be used to combine feature subset selection with holistic generalized linear models to fully automate the model selection process. Concretely, we directly optimize for the Akaike and Bayesian information criteria while imposing constraints designed to deal with multicollinearity in the feature selection task. Specifically, we propose a novel pairwise correlation constraint that combines the sign coherence constraint with ideas from classical statistical models like Ridge regression and the OSCAR model.
title Automated Model Selection for Generalized Linear Models
topic Machine Learning
Optimization and Control
68T05
G.3; G.4
url https://arxiv.org/abs/2404.16560