Incremental clustering for categorical data using clustering ensemble

More and more data in practice is changing every minute and been collected in incremental mode, and incremental clustering has attracted much of researchers´ attention. However, little research now focuses on partitioning categorical data in incremental mode. How to design incremental clustering for categorical data is an urgent problem. We propose an incremental clustering for categorical data using clustering ensemble in this paper. We firstly prune redundant attributes if needed, and then make use of true values of different attributes to form clustering memberships, and next use clustering ensemble to merge or divide clusters to gain optimal clustering. Finally, the proposed algorithm is applied in Yellow-Small dataset, Diagnosis dataset and Zoo dataset and results show that it is effective.