Framework of Clustering-Based Outlier Detection

Outlier detection is important in many fields. The concept about outlier factor of object is extended to the case of cluster. Outlier factor of cluster measure the deviation degree of a cluster from the whole dataset and two outlier factor definitions are presented. A framework of clustering-based outlier detection, named FCBOD, is presented. The framework consists of two stages, the first stage cluster dataset and the second stage determine outlier cluster by outlier factor. The time complexity of FCBOD is nearly linear with respect to both size of dataset and number of attributes. The theoretic analysis and the experimental results show that the detection approach is effective and practicable.