Robust Label Shift Quantification

Fuente: arXiv
Saved in:
Bibliographic Details
Main Author: Lecestre, Alexandre
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914320280977408
author Lecestre, Alexandre
author_facet Lecestre, Alexandre
contents In this paper, we investigate the label shift quantification problem. We propose robust estimators of the label distribution which turn out to coincide with the Maximum Likelihood Estimator. We analyze the theoretical aspects and derive deviation bounds for the proposed method, providing optimal guarantees in the well-specified case, along with notable robustness properties against outliers and contamination. Our results provide theoretical validation for empirical observations on the robustness of Maximum Likelihood Label Shift.
format Preprint
id arxiv_https___arxiv_org_abs_2502_03174
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Robust Label Shift Quantification
Lecestre, Alexandre
Statistics Theory
Machine Learning
62F35
In this paper, we investigate the label shift quantification problem. We propose robust estimators of the label distribution which turn out to coincide with the Maximum Likelihood Estimator. We analyze the theoretical aspects and derive deviation bounds for the proposed method, providing optimal guarantees in the well-specified case, along with notable robustness properties against outliers and contamination. Our results provide theoretical validation for empirical observations on the robustness of Maximum Likelihood Label Shift.
title Robust Label Shift Quantification
topic Statistics Theory
Machine Learning
62F35
url https://arxiv.org/abs/2502.03174