Aplicação de técnicas de mineração de dados à um recorte do repositório de dados históricos sobre eleições brasileiras do Tribunal Superior Eleitoral brasileiro

Abstract

This study aims to apply a set of KDD (knowledge discovery in databases) tecniques at the historical election data repository from Superior Electoral Court (Tribunal Superior Eleitoral) as a case study. To complete this goal, five classification techiniques (C4.5, Knn-3, Knn-5, Support vector Machine, Multilayer Perceptron, Naive Bayes) and one association techinique (Apriori) were chosen. For the classification experiment, the data used was about the House elections from 2010. For the association experiment, the data used was from Mayor and City Concilmans from all cities, both from the year 2012.