Ordinal regression meets online learning: Interactive preference learning for multiple criteria choice and ranking with provable guarantees

Marco Grillo, Wojciech Kotłowski and Miłosz Kadziński

European Journal of Operational Research, 2025, vol. 327, issue 3, 1003-1022

Abstract: We propose a theoretical and practical bridge between ordinal regression for multiple criteria choice and ranking problems and the framework of sequential prediction, also known as online learning. By reframing the ordinal regression as a sequential prediction task, we study a general class of algorithms that assign probabilities to a sequence of the preferences expressed by the Decision Maker (DM). This approach allows us to evaluate various statistical algorithms on a common basis, providing theoretical guarantees on their regret. To model the likelihood, we employ an additive value function that scores pairwise comparisons given by the DM. We explore two likelihood models: (1) a linear model, which we demonstrate is analogous to sequential investment, and (2) the Bradley–Terry model, widely used in statistics and preference learning. For both models, we establish theoretical bounds for the Bayesian method and the Regularized Maximum Likelihood algorithm (also known as Follow the Regularized Leader). We design Monte Carlo Markov Chain methods based on Metropolis–Hastings and Nested Sampling for efficient approximation of the posterior in Bayesian methods. Extensive empirical testing on synthetic and real-world data shows that our methods outperform the best existing approaches in the literature.

Keywords: Multiple criteria analysis; Preference learning; Online learning; Ordinal regression; Monte Carlo Markov Chain; Bayesian methods (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0377221725004436
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:ejores:v:327:y:2025:i:3:p:1003-1022

DOI: 10.1016/j.ejor.2025.05.045

Access Statistics for this article

European Journal of Operational Research is currently edited by Roman Slowinski, Jesus Artalejo, Jean-Charles. Billaut, Robert Dyson and Lorenzo Peccati

More articles in European Journal of Operational Research from Elsevier
Bibliographic data for series maintained by Catherine Liu ().