Finding Hierarchies of Subspace Clusters

Many clustering algorithms are not applicable to high-dimensional feature spaces, because the clusters often exist only in specific subspaces of the original feature space. Those clusters are also called subspace clusters. In this paper, we propose the algorithm HiSC (Hierarchical Subspace Clustering) that can detect hierarchies of nested subspace clusters, i.e. the relationships of lowerdimensional subspace clusters that are embedded within higher-dimensional subspace clusters. Several comparative experiments using synthetic and real data sets show the performance and the effectivity of HiSC.