Analysis algorithm for cluster-based outlier
基于聚类的离群点分析方法*

For the formation of outliers,different attributes play different roles.Outliers in the different attribute space will show the different characteristics.In most cases,whether the high dimensional data is the outlier usually depends on the projection of these objects in low-dimensional space.In order to classify the origin of outliers,this paper defined some concepts such as outlier attributes and outlier cluster and proposed the approaches to outlier analysis by classified outliers based on the existing outlier mining technology.Furthermore,it gave an effective CBOC algorithm to provide outliers’intensional knowledge,whose validity in practical applications is finally verified by the experimental results.