A Fuzzy K-modes-based Algorithm for Soft Subspace Clustering

This paper proposes a Fuzzy K-modes-based Algorithm for Soft Subspace Clustering, which adopts some fuzzy techniques for subspace clustering on mixed features. In order to obtain better clustering result, the proposed algorithm focuses on not only the intra-similarity of clusters, but also the optimization of the subspace where the cluster is situated. Experimental results show that the proposed FKSSC algorithm is efficient and effective in clustering both categorical and numeral data sets in high dimensional space.