APLICAÇÃO DE TÉCNICA DE MINERAÇÃO DE DADOS COMO PERSPECTIVA DE ANÁLISE DO CENÁRIO PARTIDÁRIO DAS ELEIÇÕES MUNICIPAIS DO INTERIOR GOIANO

Abstract

Currently, Brazil is a country of world reference in elections due to its modernization process by fully computerized electronic ballot boxes since 2000, thus guaranteeing the security of votes and democratic freedom. In each election, a large amount of data is generated and, based on statistical numbers, studies are carried out to build applications or discover new information. This work aims to conduct a research to apply knowledge discovery techniques in data mining, aiming to find a possible trend to be observed in the electorate profile of electoral zones 039 and 072 in the State of Goiás. With the help of the WEKA tool to perform this task, a set of data from voters and electoral parties at the municipal level, which took place between 2000 and 2016, obtained from the data repository made available by the TSE (Superior Electoral Court), was subsequently applied to Decision Tree technique using the J48 data classification algorithm, so that association rules can be generated. Finally, the results found through four experiments carried out are presented, explaining whether certain patterns in the electorate profile have an influence on the victory of a given party electoral election in the municipal elections in the areas mentioned.