POLICIAMENTO PREDITIVO E INTELIGÊNCIA ARTIFICIAL: ANÁLISE DE DESEMPENHO DO ALGORITMO DE APRENDIZADO DE MÁQUINA SUPERVISIONADO RANDOM FOREST NA PREDIÇÃO DE OCORRÊNCIAS POLICIAIS DE ROUBO NAS ZONAS DA REGIÃO METROPOLITANA DE SÃO LUÍS

Predictive policing is a concept based on the premise that it is possible to predict when and where crimes will occur again in the future by using sophisticated computer analysis based on data from previously committed crimes (UCHIDA, as cited in NORTON, 2013, pp. 32–33). The aim of this article is to analyze the spatial prediction of crimes, highlighting the importance of innovative approaches in the field of public safety management. Specifically, the article explores practical applications of using Artificial Intelligence (AI) through supervised machine learning algorithms for classification, using the emergency call records database (190) of robbery incidents in the metropolitan region of São Luís as input, and comparing results through validation techniques for the Random Forest algorithm. Through detailed analysis of these data, the potential of predictive policing is illustrated not only to anticipate criminal events but also to develop a computational tool capable of assisting and supporting strategic decision-making, significantly contributing to troop allocation management and, consequently, to the enhancement of preventive urban patrol services.