The Research of Outlier Mining Based on Constructed Covering Algorithm
基于构造性覆盖算法的离群数据挖掘研究

A new algorithm of outlier mining based on the covering algorithm is proposed in this paper. Since the par- ticularity of constructing the covering domain, the covering is very suitable for the outlier mining. After analyzing the covering model, the definition of the outlier and procedures of algorithm according the covering model is given. Through this way, the complicate problem of outlier mining is changed into very simple analyzing covering samples problem, which is intuition and give good explanation of the outlier. And the algorithm is suitable for the high dimen- sion and huge mount of data processing. Some experiments are given, and the results show the effective of the algo- rithm.