基于遗传算法的盗窃案时空关联分析

Spatio-temporal correlation analysis of theft based on genetic algorithm

针对盗窃案时空热点分布规律和形成机制的问题,采用时空关联和时空热点矩阵结合的方式进行分析。时空关联分析方法常被应用于犯罪时空关联分析中,但在规则提取过程中数据库需多次重复扫描,降低了时空关联规则提取效率,因此将模拟退火的遗传算法引入到关联分析中,通过交叉变异等方法进行规则筛选,提高时空关联规则的提取效率,并将规则引入到时空热点矩阵分析中,得到时空热点分布图,再结合归因理论、中性化理论、正强化理论等心理学理论进行成因分析,最后针对时空热点分布结果进行案件时空分布成因分析,并提出相应的防范对策。

Aiming at the problem of the distribution and formation mechanism of time-space hotspots in theft,this paper uses the combination of space-time correlation and space-time hotspot matrix to analyze.The spatio-temporal association analysis method is often applied to the spatio-temporal association analysis of crimes.However,the repeated scans of the database in the rule extraction process reduce the efficiency of spatio-temporal association rules extraction.Therefore,this study introduces the simulated annealing genetic algorithm into the correlation analysis.The rules of cross-variation are used to filter the rules,improve the extraction efficiency of spatio-temporal association rules.And the rules are introduced into the analysis of spatio-temporal hotspot matrix,to obtain the spatio-temporal hotspot distribution map.Then it combins the attribution theory,neutralization theory,positive reinforcement theory and other psychology.The paper analyzes the causes of the spatio-temporal distribution of the time-space hotspot distribution and proposes corresponding countermeasures based on the analysis of the causes of the theory.