Dimension-free uniform concentration bound for logistic regression

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: Nakakita, Shogo
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916436144816128
author Nakakita, Shogo
author_facet Nakakita, Shogo
contents We provide a novel dimension-free uniform concentration bound for the empirical risk function of constrained logistic regression. Our bound yields a milder sufficient condition for a uniform law of large numbers than conditions derived by the Rademacher complexity argument and McDiarmid's inequality. The derivation is based on the PAC-Bayes approach with second-order expansion and Rademacher-complexity-based bounds for the residual term of the expansion.
format Preprint
id arxiv_https___arxiv_org_abs_2405_18055
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Dimension-free uniform concentration bound for logistic regression
Nakakita, Shogo
Statistics Theory
Machine Learning
62J12 (Primary), 62F12 (Secondary)
We provide a novel dimension-free uniform concentration bound for the empirical risk function of constrained logistic regression. Our bound yields a milder sufficient condition for a uniform law of large numbers than conditions derived by the Rademacher complexity argument and McDiarmid's inequality. The derivation is based on the PAC-Bayes approach with second-order expansion and Rademacher-complexity-based bounds for the residual term of the expansion.
title Dimension-free uniform concentration bound for logistic regression
topic Statistics Theory
Machine Learning
62J12 (Primary), 62F12 (Secondary)
url https://arxiv.org/abs/2405.18055