两阶段孤立点挖掘算法在保险欺诈中的应用  

Two-step isolated mining algorithm based on clustering in the insurance-fraud protection

作　　者：陈亮[1] 

机构地区：[1]泰山职业技术学院信息工程系,山东泰安271000

出　　处：《长春工业大学学报》2012年第1期49-52,共4页Journal of Changchun University of Technology

摘　　要：为了识别保险业中欺诈行为,推导出了基于粗糙集的模糊集合相似性度量公式。通过改进向量相似性判断最大相异系数,提出基于聚类算法的两阶段孤立点挖掘算法,提高了算法的实时性。小范围试验结果验证了算法的有效性。To identify the fraud action in the insurance field,a similarity measurement formula based on fuzzy set is put forward.With improved vector similarity to judge the maximum alien coefficients,a two-step isolated mining algorithm based on clustering to speeding up the algorithm.The test results in a fixed range show that it is feasible.

关 键 词：集合相似性 向量相似性 聚类 孤立点 保险欺诈 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]