Abstract
This paper considers the problem of subspace clustering under noise. Specifically, we s- tudy the behavior of Sparse Subspace Clus- tering (SSC) when either adversarial or ran- dom noise is added to the unlabelled input data points, which are assumed to lie in a u- nion of low-dimensional subspaces. We show that a modified version of SSC is provably ef- fective in correctly identifying the underlying subspaces, even with noisy data. This ex- tends theoretical guarantee of this algorithm to the practical setting and provides justifi- cation to the success of SSC in a class of real applications.
Author
Yu-Xiang Wang and Huan Xu
Journal
Proceedings of the 30th International Conference on Machine Learning
Paper Publication Date
2013
Paper Type
Astroinformatics