Classification of Complaints for Criminal Intelligence Purposes

Abstract

The increase in the volume of available data is changing how people perceive their own fields and how the people may interact with this surplus of information. Public security is not different; law enforcement agencies (LEAs) now have available a large quantity of information to help them fight criminality. One challenging problem is to classify/predict criminal activities. The differentiation over two different complaints may only be clear through the careful analysis of complaints' open text fields (e.g., the modus operandi), where the specificity of the perpetrated crime is described. Sometimes the intention behind a crime is not evident unless it is correlated to other crimes and patterns get extracted from them. This chapter shows that it is possible to classify criminal data using machine learning-based methods and that open text fields, such as the modus operandi, may play a fundamental role in the performance of the classification.