关键词: 大数据分析, 离群点检测, 子空间聚类, 累积全熵

Abstract: Aiming at the instability and complex computation of optimal subspace clustering selection existed in Cumulative Mutual Information (CMI) method,the computational method for chain rule of cumulative entropy,cumulative total correlation and cumulative holoentropy were given.Cumulative holoentropy was used to mine the best clustering subspaces on continuous data sets in which outliers were detected,and then subspace outlier detection algorithm based on cumulative holoentropy was proposed.The validity and scalability of proposed method were tested on real datasets and virtual datasets.Experiment result showed that the efficiency of mining outliers in subspaces was enhanced by the proposed algorithm.

Key words: big data analysis, outlier detection, subspace clustering, cumulative holoentropy