Identificação de Indícios de Candidaturas de Fachada nas Eleições de 2018

Abstract

The proposed project seeks to identify potential fake candidates through data mining tools. At first, research was done to correlate data, such as the existence of the connection between quotas, investments in political campaigns and the success of the candidate for election. With this information collected, it was verified the existence of characteristics that presented discrepancies of the others. We then carried out research on the state of the art in the field of data mining and the algorithms that best fit the problem profile were raised. Among them, three algorithms were chosen; isolation forest that would highlight the most disadvantaged candidates, a decision tree that would highlight the profile characteristics of these candidates and k-means, which would group the candidates by their characteristics, thus tracing the profiles of potential fake candidates.