作者：单 勇
作者介绍：同济大学法学院副教授，法学博士
出版日期：2016年4月9日
内容提要 随着犯罪地理学的兴起，犯罪聚集分布及热点相对稳定分布的原因亟待释 疑。鉴于热点的空间属性，从空间相关性出发，在制图验证犯罪聚集性和热点稳定性的前 提下，以回归分析筛选空间变量，发现标志性建筑等五种因素在 63． 7% 的程度上影响盗窃 密度变化; 创建空间滞后模型，量化犯罪自相关对空间相关性的影响，发现标志性建筑对盗 窃犯罪有较强吸引力，商业点密度等因素与犯罪呈线性正相关，公交站点密度等因素与犯 罪呈非线性关系，且存在犯罪临界值。这种从相关到因果的分析思路为犯罪原因研究带来 了经验到量化、模糊到精确、定量到定性的方法论改进，从聚焦稳定热点、依靠数据决策、评 估犯罪风险、借力城市设计方面为创新立体化防控提供了理论支持。
关键词 犯罪热点区域 犯罪制图 空间相关性 犯罪临界值 立体化防控
Abstract: With the rise of criminal geography，the explanations for clumped distribution of crime and relatively stable distribution of crime hot spots need to be given urgently． Starting from the spatial relativity due to spatial attribute of the hot spots，it is found that five factors， landmarks included，affect theft density by 63． 7% on the premise of charting to verify crime cluster and hot spots’stability，which will in turn analyze and select the spatial variables． Besides，creating lag models and quantizing the influence of crime autocorrelation on spatial relativity，it is found that landmarks are more attractive to theft and that there is a positive correlation between density of shops and the number of theft，while a non-linear relation between density of bus stations and the number of theft with a crime critical value． Such analysis thought from relativity to causality improves the methodology from experience to quantization，inaccuracy to accuracy and quantitation to qualitativeness in studying causes of crime． Furthermore， it provides theoretical supports for making innovations in three-dimensional prevention and control by clustering stable hot spots，making determination in accordance with data，assessing criminal risk and leveraging urban design．