Fraud detection on payment transaction networks via graph computing and visualization

作     者：Sun Quan Tang Tao Zheng Jianbin Lin Jiale Zhao Jintao Liu Hongbao 孙权;Tang Tao;Zheng Jianbin;Lin Jiale;Zhao Jintao;Liu Hongbao

作者机构：School of Computer ScienceFudan UniversityShanghai 200433P.R.China China UnionPay Research Institute of Electronic PaymentShanghai 201201P.R.China School of Electronic Information and Electrical EngineeringShanghai Jiao Tong UniversityShanghai 200240P.R.China 

出 版 物：《High Technology Letters》 (高技术通讯（英文版）)

年 卷 期：2020年第26卷第3期

页      面：253-261页

核心收录：

学科分类：12[管理学] 02[经济学] 0202[经济学-应用经济学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 020204[经济学-金融学（含∶保险学）] 08[工学] 080402[工学-测试计量技术及仪器] 0804[工学-仪器科学与技术] 

基　　金：Supported by the National Natural Science Foundation of China(No.61972250) National Key Research and Development Program of China(No.2018AAA0100700,2016YFB1001003) the Program of Shanghai Academic/Technology Research Leader(No.19XD1433700) 

主　　题：payment fraud detection graph computing graph embedding machine learning 

摘      要：With the fast development of Internet technology,more and more payments are fulfilled by mobile Apps in an electrical way which significantly saves time and efforts for *** a change has benefited a large number of individual users as well as merchants,and a few major players for payment service have emerged in *** a result,the payment service competition becomes even fierce,and various promotion activities have been launched for attracting more users by the payment service *** this paper,the problem focused on is fraud payment detection,which in fact has been a major concern for the providers who spend a significant amount of money to popularize their payment *** paper tries the graph computing-based visualization to the behavior of transactions occuring between the individual users and ***,a network analysisbased pipeline has been *** consists of the following key components:transaction network building based on daily records aggregation;transaction network filtering based on edge and node removal;transaction network decomposition by community detection;detected transaction community *** proposed approach is verified on the real-world dataset collected from the major player in the payment market in Asia and the qualitative results show the efficiency of the method.