基于信息熵的金融网络洗钱社区发现新算法

摘要

According to the large quantity and suspicious characteristic and quantitative features of money laundering trans-action,a money laundering community discovery algorithm based on information entropy(ML-CDBIE)is proposed according to the thoughts of aggregation and indicator optimization. The characteristic of the algorithm is to discover money laundering commu-nity according to the similarity of nodes information entropy and stability of community information entropy. According to the dra-matic change of the community entropy after nodes addition,it can determine whether the nodes division is correct,or belongs to the community,which can discover the money laundering community. The experimental results show that the algorithm has high recognition rate and perfect community structure of money laundering account,and also provides a new way to discover the money laundering community of financial network.%针对洗钱交易的大数据大额可疑特征和量化特点,基于凝聚和优化指标的思想,提出一种ML-CDBIE算法.该算法根据节点信息熵的相似性和社区信息熵的稳定性进行洗钱社区发现.根据增加节点后社区熵的变化是否剧烈可以判断节点的划分是否正确,是否属于社区的成员,从而实现洗钱社区的发现.实验结果表明,该算法不仅具有洗钱账户识别率高和良好的社区结构,而且为金融网络洗钱社区发现提供了一种新途径.