Kantorovich distances between rankings with applications to rank aggregation

摘要: The goal of this paper is threefold. It first describes a novel way measuring disagreement between rankings finite set χ n ≥ 1 elements, that can be viewed as (mass transportation) Kantorovich metric, once the collection embedded in κn n× doubly-stochastic matrices. also shows such an embedding makes it possible to define natural notion median, interpreted probabilistic fashion. In addition, from computational perspective, convexification induced by approach median computation more tractable, contrast standard metric-based method generally yields NP-hard optimization problems. As illustration, methodology applied issue ranking aggregation, and shown compete with state art techniques.