A mathematical modeling approach for geographical profiling and crime prediction

Abstract

Criminal geographical profile is a methodology of investigation, which makes use of correlation between crime location and crime base of criminal to find out the most probable place of the crime base. We introduce a model to predict the future serial crime location with aid of criminal geographic profiling and the time and locations of past crime scenes. In this model, we take the effect of distance decay and the local geographic features into consideration and formulate a probability density function of the future serial crime site to realize the prediction.