Common voting rules as maximum likelihood estimators

摘要: Voting is a very general method of preference aggregation. A voting rule takes as input every voter's vote (typically, ranking the alternatives), and produces output either just winning alternative or alternatives. One potential view following. There exists "correct" outcome (winner/ranking), each corresponds to noisy perception this correct outcome. If we are given noise model, then for any vector votes, can compute maximum likelihood estimate This constitutes rule. In paper, ask following question: For which common rules does there exist model such that model? We require votes drawn independently (we show without restriction, all have property). study question both case where outcomes winners rankings. case, only some property. Moreover, sets satisfy property incomparable between two cases (satisfying in one not imply satisfying it other case).