Multi-block Bregman proximal alternating linearized minimization and its application to orthogonal nonnegative matrix factorization.
Masoud AhookhoshLe Thi Khanh HienNicolas GillisPanagiotis PatrinosPublished in: Comput. Optim. Appl. (2021)
Keyphrases
- nonnegative matrix factorization
- bregman divergences
- probabilistic latent semantic indexing
- objective function
- matrix factorization
- negative matrix factorization
- total variation
- data representation
- split bregman
- nonnegative matrix
- least squares
- spectral clustering
- principal component analysis
- document clustering
- original data
- image restoration
- sparsity constraints
- machine learning
- co occurrence
- data matrix
- semi supervised
- relational databases