An Effective Maximal Subspace Clustering Algorithm Based on Enumeration Tree

Subspace clustering is one of the best approaches for discovering meaningful clusters in high dimensional space. However, the existing algorithms often produce clusters of great redundancy that are not easy to be understood. In this paper, based on the enumeration tree of subspace, we propose a new subspace clustering algorithm MSC to find the clusters hidden in the maximal subspace. MSC uses the monotony of cluster distribution in subspace to prune the enumeration tree of subspace, and uses the simple set intersection operation of subspace to generate the clusters. Compared to the existing subspace clustering algorithm, the experimental results confirm MSC is effective for the subspace clustering.