基于Binary-SADT的可疑金融交易识别方法  

The Recognition of Suspicious Financial Transactions Based on Binary-SADT

作　　者：张成虎[1] 吴莹莹[1] 

机构地区：[1]西安交通大学经济与金融学院,陕西西安710061

出　　处：《上海金融》2012年第5期107-111,119,共5页Shanghai Finance

基　　金：国家自然科学基金项目(70771087)

摘　　要：针对目前使用静态数据挖掘技术识别可疑金融交易所面临的监测时效性低、数据覆盖面不全的问题,通过分析可疑金融交易的特征,本文提出了基于流数据分类挖掘的可疑金融交易识别算法,即Binary-SADT算法。SADT算法能够动态解决数据流挖掘中的概念漂移,Binary-SADT在SADT的基础上利用二叉排序树处理金融交易数据流中的连续属性,构建并及时更新识别可疑金融交易的分类模型。理论分析和实验结果表明该算法所构建的分类模型符合业内专家总结的可疑金融交易特征,验证了该算法的可行性和有效性。Aiming to overcome the low-timeliness and small data coverage confronted by the static data mining technology used to recognize suspicious financial transactions,through analyzing the specific features of suspicious financial transactions,the paper has proposed an algorithm for recognizing suspicious financial transactions based on stream data classification mining,namely Binary-SADT algorithm.SADT algorithm can dynamically solve concept drift in data stream mining.On this basis,Binary-SADT uses binary sort tree to process continuous attributes in the data stream of financial transactions,builds and updates the classification model to recognize suspicious financial transactions.The theoretical analysis and experimental results both show that the classification model built by Binary-SADT algorithm is consistent with the specific features of suspicious financial transactions summarized by experts,and verify the feasibility and effectiveness of the algorithm.

关 键 词：流数据分类 可疑金融交易 Binary-SADT算法 滑动窗口 

分 类 号：F830.9[经济管理—金融学]