An Empirical Study of the Perception of Criminality through Analysis of Newspapers Online

Abstract

Crime analysis represents a great challenge to law enforcement considering that the sources to use for generating intelligence are diverse in content and/or structure. However, in recent years, techniques such as natural language processing, a field of computing, artificial intelligence and linguistics have been developed that allow to study the interactions between computers and human language, and that in turn can be used effectively in the analysis of large amounts of texts and in the subsequent derivation of interesting analytical results. This paper presents a model for analysis of criminal events from online newspapers, identifying the areas with the highest crime rates through the detection and geographical mapping of critical points and the analysis of the nature of the criminal event. The evaluation of the proposed model to estimate the perception of crime in the domain of the proposed communes indicates that it is efficient in categorizing the news and the nature of these (validated by the performance indicators).