聚类与关联规则在信息舞弊识别中的应用

The Application of Clustering and Associate Rule Mining to Fraud Information Identification

针对现代电子数据迅速膨胀,传统的审计方式已经无法应对海量的业务数据,试图将数据挖掘中的聚类和关联规则算法引入审计领域.在研究聚类与关联规则算法的含义及相关算法—K-Means和Apriori算法的基础上,提出了一种基于聚类与关联规则的审计模型,并以某市城镇医疗保险的审计为例,首先利用聚类分析进行数据筛选,然后利用关联规则挖掘海量数据之间潜在的关系,为审计提供线索.文章通过案例分析为数据挖掘在信息舞弊识别领域的应用提供参考.

Considering that with the rapid expansion of electronic data, the traditional audit approachs can not cope with vast business data, this paper intend to introduce the Clustering and Association Rule Mining in the audit fields. Based on the study of the meaning of Clustering and Association Rule Mining and their Algorithm--K-Means and Apriori, this article proposed an audit model which is based on the Clustering and Association Rule Mining, at the same time, taking the audit of medical insurance of some a city as an example, it detailed first how to use the Clustering to filter data, then how to mining the potential relationships in vast data so as to determine the audit priorities and audit clues.Through the case, the article is committed to provide a reference for the application of data mining in the fraud information identification.