Criminal event modeling method based on self-excitation point process

Abstract

The invention discloses a criminal event modeling method based on a self-excitation point process, and the method comprises the steps: firstly extracting crime cases containing crime longitudes and latitudes and crime types, and forming a crime data set; extracting crime core features by using a feature selection method of a fuzzy set; extracting a core feature subset from the crime feature set; mapping crime data into a bipartite graph by using the extracted core feature subset, and dividing the bipartite graph into two parts, namely crime places and crime types; subjecting the crime types and the crime sites to embedding update, using crime intensity values in a self-excitation point process for representing crime rates of the specific crime types at given time and the crime sites, and finally, using a trigger function for representing an interaction process of criminal events, so that a dynamic process of criminal activities is modeled. Reliability and applicability of the interaction mode between the crime type and the crime places are described by adopting the self-excitation point process model, and the significance and durability characteristics of crimes and places with different potential characteristics can be found.

Description

technical field [0001] The invention belongs to the technical field of criminology, and in particular relates to a method for modeling criminal events. Background technique [0002] As a scientific method to solve the problem of increasing crime rate, crime modeling tries to dig out the internal laws of crime events and try to predict possible future crimes through the knowledge of social statistics and criminal psychology. Place. With the continuous development of social economy, the forms and means of crime are complex and changeable, but crimes always occur within a certain space. The study of crime geography from the perspective of space has become a very important aspect in crime research. On the other hand, criminology studies the randomness of criminal behavior and the diversity of crime types from different perspectives. [0003] Existing crime modeling methods can be summarized in two main ways. The first method is to model crime based on the combination of GIS ...

AI Technical Summary

Problems solved by technology

Their models all make strong parametric assumptions, which lead to missing dynamic interactions between crime events and different crime types and crime locations

[0005] Although existing crime modeling methods can predict crime within a specific city, there are two shortcomings: first, the dynamic interaction between crime and location is ignored; second, heterogeneous data is ignored. Large scale, uncertainty and ambiguity

Images