出版社：Grupo de Pesquisa Metodologias em Ensino e Aprendizagem em Ciências
摘要：Crime is a problem that governments and societies face, with several investments being made in public security and intelligence, to try to punish and prevent criminal actions. This project aims to assist in public safety, applying data mining techniques on databases containing criminal incident bulletins in the city of Divinópolis-MG. The databases have occurrences from January 2016 to May 2019, provided by the Military Police of the State of Minas Gerais (PMMG). The methodological procedures, data selection was initially performed, followed by data pre-processing and transformation. Next, data mining techniques were applied, such as: Clustering and Extraction of Association Rules. In addition, the stage was dedicated to statistical analysis related to crimes of “Theft” and “Robbery”, as well as crimes related to violence against women. Among the results, two Association Rules stand out, found using the Apriori algorithm, “Night, Robbery” => “Male Victim” and “Firearms” => “No apparent injury”, in addition to the statistical rules were performed on the data, such as “analysis of the age groups of victims” and “distribution of criminal occurrences in the week”. Thus, it is concluded that this work reached the desired goals, bringing knowledge that can be used by public security agencies. Finally, it is suggested as future works, the expansion of the database, as well as working with latitude and longitude data for each criminal occurrence.
关键词：Criminal Analysis;Data Mining in Criminal Occurrences;Extraction of association rules.
