Abstract: This paper presents an algorithm for mining the pattern similarities of outliers of sites in distributed systems. Firstly, outliers in all sites are found by the distance-based algorithm. Then knowledge sets of outliers in all sites are mined. Finally, the pattern similarities of outliers are discovered with the help of categorical and behavioral similarities of outliers of all sites.

Key words: Pattern Similarity of Outliers, Outliered Mining, Distributed Mining

摘要： 本文提出了分布式系统中各站点离群数据之间模式相似性挖掘算法，该算法首先利用基于距离的离群数据挖掘算法挖掘各自站点的离群数据，然后计算离群数据的知识集，最后依据各站点知识集的分布情况来判断离群数据之间的类别相似性和行为相似性。

关键词: 离群数据的模式相似性, 离群数据挖掘, 分布式挖掘