电信网络诈骗被害人群空间统计分布及风险特征研究--以刷单返利类为例  

Research on the Spatial Statistical Distribution and Risk Characteristics of the Telecommunications Fraud Victims--Take,for Example,the Category of Rebates on Brush Orders

作　　者：张帆锐 石拓[2] 尤慧[1] Zhang Fanrui;Shi Tuo;You Hui(Public Security Management Department,Beijing Police College)

机构地区：[1]北京警察学院 [2]北京警察学院警务情报与数据智能实验室

出　　处：《调研世界》2024年第9期78-86,共9页The World of Survey and Research

基　　金：北京警察学院大学生创新训练计划项目“电信网络诈骗被害知识图谱构建研究”(2023国创大学生项目03)的资助。

摘　　要：本文以北京市刷单返利类电信网络诈骗案件数据和被害人回访数据为样本,利用核密度估计对重点被害人分布热点空间进行识别,通过全局莫兰统计指数(GlobalMoran’sI)与局部莫兰统计指数(LocalMoran’sI)分析被害人空间分布聚集性。最后运用随机森林和Apriori算法对被害人群的易受害风险特征和被害过程进行挖掘。结果表明:北京市刷单返利类诈骗被害人群空间统计分布的热点区域在城六区,且全局空间呈现聚集现象;被害人群的被害风险与学历、年龄、职业等静态属性紧密相关,被害过程的引流、诱导、转账环节有较为频繁的规则模式。研究结论对电信网络诈骗的精准预防工作具有参考价值。This paper takes the data of telecommunications fraud cases involving returns on online single orders in Beijing and the data of victim follow-ups as the sample.Kernel density estimation is used to identify the hotspots of key victim distributions.Global Moran’s I and Local Moran’s I are used to analyze the spatial distribution agglomeration of the victims.Finally,random forest and Apriori algorithms are employed to mine the risk characteristics of the victim population and the victimization process.The results show that:the hotspot areas of telecommunications fraud cases involving returns on online single orders in Beijing are in the six urban districts,and there is a gathering phenomenon in the overall space.The risk of victimization among the victim population is closely related to static attributes such as education level,age,and occupation.There are frequent regular patterns in the stages of victimization process such as attraction,inducement,and money transfer.The research conclusions have reference value for the precision prevention of telecommunications fraud.

关 键 词：电信网络诈骗 时空统计分布 随机森林 统计核密度估计 APRIORI 

分 类 号：C913.9[经济管理]