摘要
目前,反洗钱成为世界各政府机构关注的热点领域。本文从技术角度探讨了外汇领域的洗钱侦测系统及其关键算法的实现。首先,描述了我国第一个外汇反洗钱侦测系统的架构;然后提出了一个以语义核心树SCT(Seman-tic Core Tree)为基础的增量概念聚类算法。该算法能解决以下问题:1)能处理海量数据集;2)能处理分类和数值型混合的数据集;3)能够清楚地解释聚类结果,使得结果易于理解。该算法已在反洗钱框架下实现并投入使用。
At present, anti-money laundering has become a hot topic of the entire world. From the point of view of technique, this paper discusses a money-laundering detection system for administration of foreign exchange and one of its key algorithms. It first gives the framework of the system, which is the first anti-money laundering system of our nation. And then it proposes an SCT（Semantic Core Tree）-based incremental conceptual clustering algorithm. The al- gorithm could solve the problems of 1） large volume of data set; 2） mixture of categorical and numerical data; 3） easy understanding of result. The algorithm has been employed in the money laundering detection system .
出处
《计算机科学》
CSCD
北大核心
2007年第3期201-204,共4页
Computer Science
基金
国家自然科学基金(60403027)
关键词
反洗钱系统
数据挖掘
增量概念聚类
语义核心树
分类属性
Anti-money laundering system, Data mining, Incremental conceptual clustering, Semantic core tree, Categorical
