基于社会网络分析法的洗钱犯罪数据挖掘侦查技术的改进  

Improvement of Investigation Techniques in Money Laundering Crime's Data Mining Based on SNA

作　　者：牛惊雷 牛易航 Niu Jinglei;Niu Yihang(China People's Police University,Langfang 065000,China;Hexi Public Security Sub-Bureau of Tianjin,Tianjin 300202,China)

机构地区：[1]中国人民警察大学,河北廊坊065000 [2]天津市公安局河西区公安分局,天津300202

出　　处：《贵州警察学院学报》2024年第6期71-78,共8页Journal of Guizhou Police College

摘　　要：随着互联网信息技术的发展,洗钱犯罪手段、方式愈发复杂和隐蔽,极大地影响了可疑交易识别精度,使侦查工作陷入低效的困境,进一步改进洗钱犯罪数据挖掘技术是必然趋势。“结构二重性”理论拓展侦查视角,社会网络分析法揭示了可疑交易的识别不仅需要结合大量账户信息和交易特征,更需要深度考量与账户交易关联的网络结构特征。图卷积神经网络理论作为能够深度融合基础特征与社会网络结构特征的学习算法,为洗钱犯罪数据挖掘侦查技术提供了改进途径。The development of Internet information technology has made the criminal means and methods of money laundering crime increasingly complex and concealed,therefore substantially impairing the accuracy of suspicious transaction identification and plunging the investigation work into an inefficient dilemma.Therefore,the enhancement of the data mining techniques for the detection of money laundering crime has become an inexorable trend.The"duality of structure"theory helps expand the perspective of investigation,while social network analysis reveals that the identification of suspicious transactions requires not only extensive account information and transaction features,but also an in-depth examination of the network structure associated with account transactions.The theory of Graph Convolution Network,or GCN,as a learning algorithm capable of deeply integrating basic features with social network structure attributes,provides an approach for improving investigation techniques applied in money laundering crime's data mining.

关 键 词：洗钱犯罪 可疑交易 社会网络 图卷积神经网络 

分 类 号：D917.6[政治法律—法学]