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.
出处
《长春工业大学学报》
CAS
2012年第1期49-52,共4页
Journal of Changchun University of Technology
关键词
集合相似性
向量相似性
聚类
孤立点
保险欺诈
set similarity
vector similarity
clustering
lsolated-point
insurance-fraud.
