多尺度时空特征约束的犯罪预测方法—以入室盗窃为例

项目名称： 多尺度时空特征约束的犯罪预测方法—以入室盗窃为例

项目编号： No.41501488

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 天文学、地球科学

项目作者： 王增利

作者单位： 南京林业大学

项目金额： 20万元

中文摘要： 多尺度犯罪预测是在分析犯罪多个时空尺度特征的基础上量化犯罪的分布规律，实现对犯罪的预测。目前对于犯罪分析和预测的研究大多在单一尺度上进行。本项目将以入室盗窃为例，在多个时空尺度上分析犯罪的关联因子，揭示地理因子与犯罪的关联关系；研究犯罪的时空分布规律，量化犯罪“聚集-转移-聚集”的时空过程；挖掘时空犯罪过程的动态特征，基于分子“吸引-排斥”的原理构建犯罪时空预测模型。研究将犯罪分析从单一尺度推向多个尺度的层面，从数学模型预测推向机理模型预测的层面。研究不仅能加强人们对于犯罪地理机理和时空规律的认识，丰富犯罪地理学的理论和方法，而且在公安实战和社会安全方面具有较高的实用价值。

中文关键词： GIS；多尺度；时空特征；犯罪；预测

英文摘要： In order to improve the performance of crime prediction, multi-resolution crime forecasting is proposed based on analysis of multi-scale spatial-temporal characteristics and pattern of crime’s distribution. Most researches are conducted on a single resolution at present. This project will be conducted with burglary data. The relationship between geographical characteristics and burglary will be revealed by the multi-resolution analysis using regression model. By quantizing the spatial-temporal process of burglary, the dynamic pattern of “gathering-transfer-gathering” will be revealed. A burglary forecasting model will be proposed based on the theory of intermolecular forces: “attraction-repulsion” to simulate the dynamic pattern mined. Through this research, the single-resolution crime analysis will be promoted to a higher level of multi-resolution analysis. What is more, the mathematical model will also be advanced to a mechanism model with higher accuracy. In conclusion, our work will not only deepen the understanding of the mechanism and pattern of burglary, rich the theory and method of environmental criminology, but also has great theoretical and practical meanings for real public security and social security.

英文关键词： GIS;multi-resolution;spatial-temporal characteristic;crime;prediction