PRINCIPAL COMPONENTS ANALYSIS AS A TOOL TO IDENTIFY VULNERABILITY TO ATM EXPLOSION ATTACKS IN THE MIDWEST OF MINAS GERAIS

Abstract

Attacks at ATMs are a reality in Brazil. They have become one of the criminal activities with the greatest social repercussion and it challenges the reactive capacity of the organs responsible for public security, since, mainly, the phenomenon presents itself through the internalization of violence, reaching less structured municipalities. Considering the difficulty of predicting how this criminal modality will behave, spatially, there is a great challenge to act in a preventive way. Considering the economic perspective, there are several variables that can influence the decision of the criminal in choosing a target as to the possibilities of gains and risk of imprisonment or punishment. Thus, the central objective of this work is to define a set of variables that combined to allow an analysis of the municipalities most vulnerable to the criminal incidence of ATM attacks. For this purpose, we used the Principal Components Analysis (PCA), a multivariate statistical analysis technique, aimed at the synthesis of a large number of variables in a single factor that comprises a considerable portion of the data variance. The spatial analysis of the work took place in the 7th Military Police Region and analyzed events from the period of 2012 to 2017. The results allowed us to describe the events triggered, to verify the correlation between the variables listed to characterize the municipalities and to analyze the vulnerability of attacks.