Robust binary and multinomial logit models for classification with data uncertainties

Baichuan Mo

doi:10.1016/j.ejor.2025.05.013

Robust binary and multinomial logit models for classification with data uncertainties

2025-05-22

0

PRE

AI

DOI：10.1016/j.ejor.2025.05.013

Original

Original request for help

Save

Abstract

En 中文

• Reformulate DCMs with robust optimization for classification under test errors. • Derive robust models for binary/multinomial logits with feature and label uncertainty. • Derive statistical properties of robust estimators. • Experiments show better model accuracy and generalizability.

Keywords:

robust optimization

classification

feature uncertainty

label uncertainty

statistical properties

Robust binary and multinomial logit models for classification with data uncertainties

Abstract

Journal

Organization