The Open Cybernetics & Systemics Journal

2014, 8 : 860-868
Published online 2014 December 31. DOI: 10.2174/1874110X01408010860
Publisher ID: TOCSJ-8-860

Research of Crime Prediction Technology Based on Mathematical Model

Zhu Ke and Zhang Jin
Department of Educational Technology, Henan Normal University, Xinxiang, Henan, 453002, China.

ABSTRACT

The main tasks of this study are to ascertain the geographic profile of the criminal to find out the anchor point based on the time and locations of previous crimes and to predict the potential locations of the next crime. Firstly, this paper utilizes probability and statistics method to calculate the possibility of every point becoming the criminal’s anchor point, from which this paper develops two schemes—Distance Function Method and Distribution Function Method to generate the geographic profile. Then with the result of the Distance Function Method as a reference point, the criminal’s anchor point could be pinpointed from the result of the Distribution Function Method. Secondly, by Multivariate Analysis Method, this paper defines the Euclidean and Manhattan Distances between the anchor point and the locations of the previous crime sites, and then, according to distribution features of these distances, selects the corresponding distribution function to get the concrete expression. Thirdly, this paper adopts Fuzzy Mathematical Method to get quantified and normalized index factors and use Analytic Hierarchy Process to compute different weights of social index factors of different areas in the region. Thus, the weight scores of every area are gotten. Finally, considering the natural factors and social factors simultaneously, this paper determines the final probability distribution of the locations of the next crime, by which this paper could predict the most likely location of the next crime.

Keywords:

Data mining, mathematical modeling, crime prediction.