团伙犯罪中基于PageRank的嫌疑人犯罪影响力分析  

Analysis of the influence of suspects in gang crimes based on PageRank

作　　者：张俊豪 Zhang Junhao(Department of Image and Network Investigation,Zhengzhou Police College,Zhengzhou 450003,China)

机构地区：[1]郑州警察学院图像与网络侦查系,郑州450003

出　　处：《河南师范大学学报（自然科学版）》2024年第2期81-88,共8页Journal of Henan Normal University(Natural Science Edition)

基　　金：河南省科技攻关项目(212102210531);河南省高等学校重点科研项目计划(23B630017);郑州警察学院2023年基科费项目(2023TJJBKY038).

摘　　要：如何准确评估嫌疑人的影响力是侦破团伙犯罪的关键,为此,在PageRank算法的基础上,结合团伙犯罪的网络结构特性,提出团伙犯罪中嫌疑人犯罪影响力评估模型.该模型首先结合已掌握的作案信息将团伙犯罪网络转化为有向加权网络,再根据中间中心度、接近中心度两大网络特性确定罪犯影响力转移矩阵,最后该模型以实际的案例为实验数据得出具体的嫌疑人犯罪影响力,并以两种成熟的模型为对比模型,实现结果显示该模型的评估结果与实际案件结果拟合度较高,且准确率也高于对比模型,验证了该模型具有较高的准确性和可行性.How to accurately evaluate the influence of suspects is the key to detecting gang crime.Therefore,based on the PageRank algorithm and the network structure characteristics of gang crime,a model for evaluating the influence of suspects in gang crime is proposed.The model first converts the gang criminal network into a directed-weighted network based on the known crime information,and then determines the criminal influence transfer matrix according to the two network characteristics of middle centrality and proximity centrality.Finally,the model uses actual cases as experimental data to obtain specific suspects'criminal influence,and uses two mature models as comparison models,The implementation results show that the evaluation results of the model have a high degree of fitting with the actual case results,and the accuracy is also higher than the comparison models,verifying the high accuracy and feasibility of the model.

关 键 词：PAGERANK 网络结构特性 团伙犯罪 嫌疑人影响力 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]