A review on applied statistical and artificial intelligence techniques in crime forecasting

Abstract

Crime forecasting is an important component of crime analysis towards providing early information about possible crime occurrences in the future. Different models have been proposed to assess different crime data structures and representations. From the literature study conducted, there are several types of crime forecasting models that have been introduced such as statistical model and artificial intelligence (AI) model. Recent trends indicate that researchers have shifted their interest towards AI model due to its flexibility in handling variations in crime data structures. The study found that AI model is capable of capturing nonlinearity pattern of crime data in which statistical model fails to achieve. Moreover, the structure of crime data is mostly nonlinear. Thus, an AI model is favoured among researchers towards developing a robust crime forecasting model. This paper provides a review on the background, trends, and challenges on applied statistical and AI model in crime forecasting.