摘要
传统的犯罪地理目标模型仅基于案件点空间距离信息进行预测,忽视了地理环境因素、警察因素等对犯罪人选择落脚点的影响,导致预测精度不理想。针对这一问题,将地理环境、警察等高相关性的因素通过模糊化处理提取出来,作为模型修正参数加入到传统CGT模型中,从而得到改进的CGT模型。利用实例对传统模型与改进模型预测效果进行对比,结果表明改进CGT模型取得了较好的预测效果,预测命中百分率至少高于原始CGT模型6%以上。
The traditional geographic profiling model is only based on time and spatial distance of cases,and ignores the factors like geographical environment and police,etc.in determining criminals' choice of foothold,which leads to lowaccuracy of prediction.To address this,high correlation factors such as geographical environment and police are extracted as model correction parameters by the fuzzy processing to be added to the traditional CGT model.The improved CGT model is tested on cases to compare the results with the traditional model.The results showthe improved model achieves good prediction,and the hit rate increases by at least 6 percentage points.
出处
《常州工学院学报》
2017年第4期28-32,共5页
Journal of Changzhou Institute of Technology
基金
公安部技术研究计划重点项目(2016JSYJA17)
关键词
犯罪地理画像
犯罪地理目标模型
模糊算法
地理环境因素
警察因素
criminal geographic portrait
crime geographic targeting model
fuzzy algorithm
geographical environment factors
police factors