基于空间分析模型的犯罪率与个人生活指标相关性分析——以美国50个州及哥伦比亚特区为例

Correlation Analysis of Crime Rate and Personal Life Indicators Based on Spatial Analysis Model:Take the 50 U.S. States and the District of Columbia as Examples

犯罪行为是威胁社会安定的关键因素。本文以美国50个州及哥伦比亚特区为研究对象,选取青壮年男子受教育状况、私人交通及通信状况等3种个人生活指标,探究不同空间分析模型下,犯罪率与上述指标之间的关系,进一步分析各个指标与犯罪率之间的相关性,有望对区域治安管理与犯罪行为分析提供一定的参考。结果显示,犯罪率具有空间自相关的特性,相较于OLS模型、空间自相关模型、空间滞后模型和空间误差模型,地理加权回归模型更适于进行犯罪率与影响因素的空间相关性分析;受教育程度低的青壮年男子占比与犯罪率呈现正相关关系,移动通信设备占有率和私人交通工具占有率对犯罪率在不同地区具有不同的影响。

Criminal behavior is a key factor threatening social stability.This article takes 50 U.S.states and the District of Columbia as the research objects,and selects three personal life indicators including young men’s education status,the occupancy of private transportation and communication,to explore the applicability of different spatial analysis models on the correlation between crime rates and the above indicators,and further analyzes the correlation between the indicators and crime rates,which is expected to provide reference value for regional security management and criminal behavior analysis.The results show that the crime rate has the characteristics of spatial autocorrelation.Compared with the OLS model,spatial autocorrelation model,spatial lag model and spatial error model,the geographically weighted regression model is more suitable for analyzing the spatial correlation between crime rate and indicators.At the same time,the proportion of young men with low education level is positively correlated with crime rate,and the occupancy rate of mobile communication equipment and the occupancy rate of private transportation has different effects on crime rate in different regions.