基于图分析算法的信用卡交易欺诈检测  

Credit Card Transaction Fraud Detection Based on Graph Analysis Algorithm

作　　者：袁立宁 唐雨霞 黄琬雁 罗恒雨 何佩遥 YUAN Lining;TANG Yuxia;HUANG Wanyan;LUO Hengyu;HE Peiyao(School of Information Technology,Guangxi Police College,Nanning 530028,China;School of Public Security Big Data Modern Industry,Guangxi Police College,Nanning 530028,China;School of Criminal Science and Technology,Guangxi Police College,Nanning 530028,China)

机构地区：[1]广西警察学院信息技术学院,广西南宁530028 [2]广西警察学院公安大数据现代产业学院,广西南宁530028 [3]广西警察学院刑事科学技术学院,广西南宁530028

出　　处：《现代信息科技》2024年第15期138-141,共4页Modern Information Technology

基　　金：广西高校中青年教师科研基础能力提升项目(2024KY0902);广西哲学社会科学研究课题(23FTQ005);广西壮族自治区公安厅专项课题(2023GAQN092,2023GAQN110)。

摘　　要：当前,在线信用卡交易欺诈案件快速增加,作案手段和方法更加多变,信用卡交易欺诈检测已成为银行风险防控的重点内容。文章依托近年人工智能领域热门的图分析理论与算法,将信用卡交易数据转化为图结构数据,从而分析信用卡交易欺诈图的社区信息。在此基础上,应用图表示学习算法Deepwalk和机器学习分类器,构建信用卡交易欺诈检测模型,用于预测欺诈行为。实验结果表示,该模型对欺诈行为的检测准确率达70%。Currently,online credit card transaction fraud cases are rapidly increasing,with more diverse methods and tactics being used.The credit card transaction fraud detection has become the key focus of bank risk prevention and control.This paper relies on the popular graph analysis theory and algorithms in the field of Artificial Intelligence in recent years.It transforms credit card transaction data into graph-structured data to analyze the community information of the credit card transaction fraud graph.Based on this,it applies the graph representation learning algorithm Deepwalk and Machine Learning classifiers,and a credit card transaction fraud detection model is constructed to predict fraudulent behavior.The experiment results show that the model detection accuracy for fraud behavior reaches 70%.

关 键 词：信用卡交易 欺诈检测 图分析算法 图表示学习 

分 类 号：F832.2[经济管理—金融学] TP18[自动化与计算机技术—控制理论与控制工程]