关键词: 聚类, 距离, 离群数据

Abstract: The identifying, analyzing and evaluating algorithm of finding distance-based Outlier (Cell-Based) was firstly studied, and its advantages and disadvantages were pointed out. And then, a new Outlier finding algorithm-algorithm of finding Outlier for reclustering based on distance was proposed. Theoretical analysis and experimental results show that this algorithm can not only effectively overcome the faults of traditional Cell-Based algorithm, i.e. need to be recomputed from scratch for every change of the parameters, and only suitable for finding the Outlier of low dimension, but also obviously avoid the problems caused by randomly selecting initial value to produce different finding results of Outlier at higher convergence speed.

Key words: cluster, distance, Outlier