Kernel sparse subspace clustering

Abstract

Subspace clustering refers to the problem of grouping data points that lie in a union of low-dimensional subspaces. One successful approach for solving this problem is sparse subspace clustering, which is based on a sparse representation of the data. In this paper, we extend SSC to non-linear manifolds by using the kernel trick. We show that the alternating direction method of multipliers can be used to efficiently find kernel sparse representations. Various experiments on synthetic as well real datasets show that non-linear mappings lead to sparse representation that give better clustering results than state-of-the-art methods.