Aplicación de minería descriptiva de datos para el descubrimiento y visualización de patrones en delitos de robo en Ecuador.

Abstract

This paper addresses the growing problem of crime, specifically theft crimes in Ecuador, by presenting an approach that uses descriptive data mining techniques to manage information on theft crimes provided by the Fiscalía General del Estado. These techniques seek to identify patterns and visualize useful information from the data, with the objective of identifying related theft crime news. The study focuses on the use of clustering algorithms to identify relevant geographic zones in each province, followed by the extraction of association rules to find theft crime patterns. It is concluded that the CRISP-DM methodology is effective, as well as the use of K-means to identify significant geographical zones in each province and FP-growth for association rule mining. In addition, the optimal minimum support for the extraction of association rules, which derive theft patterns, which will then relate crime news, is determined. The results obtained allow the identification of crime patterns and related news, contributing to a better understanding and management of crime in the country.