Research on Analysis Technology of Burglary Point Pattern Based on LGCP Model

Abstract

Burglary is a violent criminal act that infringes on the property of others. In order to better understand the criminal model and make full use of law enforcement resources, this paper uses the space-time analysis technology to study the spatial distribution, spatial dependence and the rules of burglary cases in the main urban area of a city in 2016. The LGCP (Log-Gaussian Cox Process) model combines with the Bayesian framework to establish a continuous spatial random fields. The model is calculated using INLA(Integrated Nested Laplace Approximation) and SPDE(Stochastic Partial Differential Equations). Studies have shown that burglary cases have different trend values at different time points and different spatial locations, and have strong correlations in adjacent time periods and spatial random fields. The intensity of burglary and the spatial distribution of retail points and cultural education are linear relationship, and there is a complex nonlinear relationship with the spatial trend of sports and leisure points, but it presents a positive correlation in the whole, the more concentrated the sports and leisure, cultural education and wholesale and retail are, the higher the risk of burglary.