Saved in:
Bibliographic Details
Main Authors: Chen, Gengyang, Zhu, Mu
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2510.25632
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916026701053952
author Chen, Gengyang
Zhu, Mu
author_facet Chen, Gengyang
Zhu, Mu
contents We extend a heuristic method for automatic dimensionality selection, which maximizes a profile likelihood to identify "elbows" in scree plots. Our extension enables researchers to make automatic choices of multiple hyper-parameters simultaneously. To facilitate our extension to multi-dimensions, we propose a "softened" profile likelihood. We present two distinct parameterizations of our solution and demonstrate our approach on elastic nets, support vector machines, and neural networks. We also report a small simulation study to investigate violations to an assumption we make, and briefly discuss applications of our method to other data-analytic tasks than hyper-parameter selection.
format Preprint
id arxiv_https___arxiv_org_abs_2510_25632
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Automatic selection of hyper-parameters via the use of softened profile likelihood
Chen, Gengyang
Zhu, Mu
Methodology
We extend a heuristic method for automatic dimensionality selection, which maximizes a profile likelihood to identify "elbows" in scree plots. Our extension enables researchers to make automatic choices of multiple hyper-parameters simultaneously. To facilitate our extension to multi-dimensions, we propose a "softened" profile likelihood. We present two distinct parameterizations of our solution and demonstrate our approach on elastic nets, support vector machines, and neural networks. We also report a small simulation study to investigate violations to an assumption we make, and briefly discuss applications of our method to other data-analytic tasks than hyper-parameter selection.
title Automatic selection of hyper-parameters via the use of softened profile likelihood
topic Methodology
url https://arxiv.org/abs/2510.25632