Stochastically Transitive Models for Pairwise Comparisons: Statistical and Computational Issues
From MaRDI portal
Abstract: There are various parametric models for analyzing pairwise comparison data, including the Bradley-Terry-Luce (BTL) and Thurstone models, but their reliance on strong parametric assumptions is limiting. In this work, we study a flexible model for pairwise comparisons, under which the probabilities of outcomes are required only to satisfy a natural form of stochastic transitivity. This class includes parametric models including the BTL and Thurstone models as special cases, but is considerably more general. We provide various examples of models in this broader stochastically transitive class for which classical parametric models provide poor fits. Despite this greater flexibility, we show that the matrix of probabilities can be estimated at the same rate as in standard parametric models. On the other hand, unlike in the BTL and Thurstone models, computing the minimax-optimal estimator in the stochastically transitive model is non-trivial, and we explore various computationally tractable alternatives. We show that a simple singular value thresholding algorithm is statistically consistent but does not achieve the minimax rate. We then propose and study algorithms that achieve the minimax rate over interesting sub-classes of the full stochastically transitive class. We complement our theoretical results with thorough numerical simulations.
Cited in
(36)- Adaptive risk bounds in unimodal regression
- Optimal rates of statistical seriation
- Editorial: Special issue on ``Nonparametric inference under shape constraints
- Nonparametric shape-restricted regression
- Stochastic transitivity: axioms and models
- Towards optimal estimation of bivariate isotonic matrices with unknown permutations
- On the estimation of latent distances using graph distances
- Isotonic regression with unknown permutations: statistics, computation and adaptation
- To stay discovered: on tournament mean score sequences and the Bradley-Terry model
- Optimal full ranking from pairwise comparisons
- Worst-case versus average-case design for estimation from partial pairwise comparisons
- Estimation of Monge matrices
- Active ranking from pairwise comparisons and when parametric assumptions do not help
- Consistency and consensus modeling of linear uncertain preference relations
- Generalized transitivity: a systematic comparison of concepts with an application to preferences in the Babington Smith model
- Spectral method and regularized MLE are both optimal for top-\(K\) ranking
- Adaptation in multivariate log-concave density estimation
- Generalized rank-breaking: computational and statistical tradeoffs
- Learning from comparisons and choices
- Design and analysis of the NIPS 2016 review process
- Simple, Robust and Optimal Ranking from Pairwise Comparisons
- Maximum selection and sorting with adversarial comparators
- A new and flexible approach to the analysis of paired comparison data
- Preference-based online learning with dueling bandits: a survey
- scientific article; zbMATH DE number 7306865 (Why is no real title available?)
- Low permutation-rank matrices: structural properties and noisy completion
- A discrete probabilistic model for analyzing pairwise comparison matrices
- Asymptotically Optimal Sequential Design for Rank Aggregation
- Re-thinking high-dimensional mathematical statistics. Abstracts from the workshop held May 15--21, 2022
- Optimal permutation estimation in crowdsourcing problems
- A General Pairwise Comparison Model for Extremely Sparse Networks
- Localization in 1D non-parametric latent space models from pairwise affinities
- A cross-validation framework for signal denoising with applications to trend filtering, dyadic CART and beyond
- Stochastic comparisons for time transformed exponential models
- Debiasing evaluations that are biased by evaluations
- Optimal level set estimation for non-parametric tournament and crowdsourcing problems
This page was built for publication: Stochastically Transitive Models for Pairwise Comparisons: Statistical and Computational Issues
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2989634)
