Ranking in the generalized Bradley-Terry models when the strong connection condition fails

DOI10.1080/03610926.2013.809114MaRDI QIDQ2807685zbMATH OpenOpenAlexFDO

Authors Ting Yan

Publication date 25 May 2016

Published in Communications in Statistics. Theory and Methods (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1411.1168

zbMATH Keywords

paired comparisonsrankinggeneralized Bradley-Terry-modelspenalized maximum likelihood estimate

Mathematics Subject Classification ID

Statistical ranking and selection procedures (62F07) Paired and multiple comparisons; multiple testing (62J15)

Abstract: For nonbalanced paired comparisons, a wide variety of ranking methods have been proposed. One of the best popular methods is the Bradley-Terry model in which the ranking of a set of objects is decided by the maximum likelihood estimates (MLEs) of merits parameters. However, the existence of MLE for the Bradley-Terry model and its generalized models to allow for tied observation or home-field advantage or both to occur, crucially depends on the strong connection condition on the directed graph constructed by a win-loss matrix. When this condition fails, the MLE does not exist and hence there is no solution of ranking. In this paper, we propose an improved version of the $varepsilon$ singular perturbation proposed by Conner and Grant (2000), to address this problem and extend it to the generalized Bradley-Terry models. Some necessary and sufficient conditions for the existence and uniqueness of the penalized MLEs for these generalized Bradley-Terry-$varepsilon$ models are derived. Numerical studies show that the ranking is robust to the different $varepsilon$. We apply the proposed methods to the data of the 2008 NFL regular season.

Recommendations

• Fitting Bradley Terry models using a multiplicative algorithm
• MM algorithms for generalized Bradley-Terry models.
• Asymptotics when the number of parameters tends to infinity in the Bradley-Terry model for paired comparisons
• On a connection between the Bradley-Terry model and the Cox proportional hazards model
• An estimation of generalized Bradley-Terry models based on the EM algorithm

Cites work

• scientific article; zbMATH DE number 41025 (Why is no real title available?)
• scientific article; zbMATH DE number 48904 (Why is no real title available?)
• A Bayesian approach to paired comparison experimentation
• An extension of Zermelo's model for ranking by paired comparisons
• Asymptotics when the number of parameters tends to infinity in the Bradley-Terry model for paired comparisons
• MM algorithms for generalized Bradley-Terry models.
• Pairwise Comparison and Ranking in Tournaments
• Ranking from unbalanced paired-comparison data
• Solution of a Ranking Problem from Binary Comparisons
• Sparse paired comparisons in the Bradley-Terry model
• The Perron–Frobenius Theorem and the Ranking of Football Teams

Cited in

(5)

• scientific article; zbMATH DE number 123309 (Why is no real title available?)
• An extension of Zermelo's model for ranking by paired comparisons
• On a connection between the Bradley-Terry model and the Cox proportional hazards model
• A note on ranking in the Plackett-Luce model for multiple comparisons
• The Bradley-Terry condition is \(L_1\)-testable

This page was built for publication: Ranking in the generalized Bradley-Terry models when the strong connection condition fails

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2807685)