Large-scale non-negative subspace clustering based on Nyström approximation

Hongjie Jia, Qize Ren, Longxia Huang, Qirong Mao, Liangjun Wang, Heping Song

Abstract: Highlights•A novel large-scale non-negative subspace clustering model based on Nyström approximation is designed.•The objective function of anchor affinity matrix considers both the local and global data characteristics.•The update rule of anchor affinity matrix is derived according to the non-negative Lagrangian relaxation.•Two effective Laplacian matrix decomposition methods based on Nyström approximation are developed.•The effectiveness of the proposed model is demonstrated by theoretical analysis and extensive experiments.