Bayesian Estimators As Voting Rules
نویسنده
چکیده
We investigate the fairness of Bayesian estimators (BEs) by viewing them as (irresolute) voting rules and evaluating their satisfaction of desirable social choice axioms. We characterize the class of BEs that satisfy neutrality by the class of BEs with a neutral structure. We prove that a BE with neutral structure is a minimax rule if it further satisfies parameter connectivity. We prove that no BE satisfies strict Condorcet criterion. We also propose three new BEs of natural frameworks and investigate their satisfaction of monotonicity and Condorcet criterion, and computational complexity.
منابع مشابه
A game-theoretic framework for classifier ensembles using weighted majority voting with local accuracy estimates
In this paper, a novel approach for the optimal combination of binary classifiers is proposed. The classifier combination problem is approached from a Game Theory perspective. The proposed framework of adapted weighted majority rules (WMR) is tested against common rank-based, Bayesian and simple majority models, as well as two soft-output averaging rules. Experiments with ensembles of Support V...
متن کاملMax-Margin Majority Voting for Learning from Crowds
Learning-from-crowds aims to design proper aggregation strategies to infer the unknown true labels from the noisy labels provided by ordinary web workers. This paper presents max-margin majority voting (MV) to improve the discriminative ability of majority voting and further presents a Bayesian generalization to incorporate the flexibility of generative methods on modeling noisy observations wi...
متن کاملCommon Voting Rules as Maximum Likelihood Estimators
Voting is a very general method of preference aggregation. A voting rule takes as input every voter’s vote (typically, a ranking of the alternatives), and produces as output either just the winning alternative or a ranking of the alternatives. One potential view of voting is the following. There exists a “correct” outcome (winner/ranking), and each voter’s vote corresponds to a noisy perception...
متن کاملTruth - tracking judgment aggregation over interconnected issues ∗
This paper analyses the problem of aggregating judgments over multiple interconnected issues. Voters share a common preference for reaching true collective judgments, but hold private information about what the truth might be. Information conflicts may occur both between and within voters. Following Bozbay, Dietrich and Peters (2014), we assume strategic voting in a Bayesian voting game setting...
متن کاملIncentive-Compatible Voting Rules with Positively Correlated Beliefs
We study the consequences of positive correlation of beliefs in the design of voting rules in a model with an arbitrary number of voters. We propose a notion of positive correlation, based on the likelihood of agreement of the k best alternatives (for any k) of two orders called TS correlation. We characterize the set of Ordinally Bayesian Incentive-Compatible (OBIC) (d’Aspremont and Peleg (198...
متن کامل