B0508
Title: An application of depth functions for social choice theory
Authors: Jean-Baptiste Aubin - Insa-Lyon (France) [presenting]
Antoine Rolland - ERIC (France)
Enguerrand Brun - ENS-Lyon (France)
Samuela Leoni - INSA-Lyon (France)
Irene Gannaz - INP-Grenoble (France)
Abstract: Recent voting methods are based on evaluations of candidates given by the voters (e.g. range voting, approval voting, majority judgment). In this case, voters can be visualised with respect to their evaluations in the space of the candidates. Given this scatterplot and a depth function, the deepest voting consists of electing the favourite candidate of the deepest point of the considered depth function of the scatterplot of the evaluations of the voters. Nevertheless, classical voting methods are based on rankings of candidates by voters (e.g. Borda's voting, Condorcet's voting, plurality, etc.). New members of the deepest voting family are investigated, based on depth functions on rankings. It is shown that deepest voting formalism unifies a large class of classical voting methods and allows the introduction of numerous new voting methods. Moreover, links between properties of depth functions and those of their associated voting method are studied.
