Recent Research Trends on Geographical Crime Prediction

Abstract

Geographical crime prediction have been the focus of much research in western countries over the past decade and crime prediction systems are already in use several countries including Japan where a prefectural police department have recently introduced a certain system. However there has been no prior research into this field in Japan. This paper presents a systematic review of geographical crime prediction and discusses their relevance to the Japanese context. We identify four categories of geographical crime prediction methods: (1) surveillance of space-time clusters of crime; (2) estimation of crime intensity based on space-time interaction; (3) prediction of crime risk based on environmental factors; and (4) prediction of crime numbers/possibilities. These categories are based on established theories and have been developed independently of each other. Finally, we suggest directions for future developments of this research field in Japan.