Applying Crime Prediction Techniques to Japan: Comparison Between Risk Terrain Modeling and Other Methods

• In Event: New Directions in Statistical Models to Study Neighborhood Crime

Abstract

In the recent times, the field of crime prediction is drawing attention in Japan. However, predicting crime is especially difficult in Japan since its crime rate is considerably lower than other developed countries and the frequency of crime is insignificant to develop a statistical model for crime prediction. Since Risk Terrain Modeling (RTM) is a method that mainly depends on the environmental factors of crime and does not require past crime data, it could be suitable for Japan. In this study, we will apply RTM to case of theft from vehicles in Fukuoka and evaluate its predictive performance in comparison with other crime prediction techniques. Then, we will discuss the merits and faults of using RTM for Japan. Based on the results, we will suggest a developed model of RTM for the Japanese situation by considering past patterns and seasonality of crimes.

Authors

• Tomoya Ohyama, University of Tsukuba

• Mamoru Amemiya, University of Tsukuba