Using underapproximations for sparse nonnegative matrix factorization.
Nicolas GillisFrançois GlineurPublished in: Pattern Recognit. (2010)
Keyphrases
- high dimensional
- nonnegative matrix factorization
- negative matrix factorization
- probabilistic latent semantic indexing
- low dimensional
- sparsity constraints
- active set
- data representation
- matrix factorization
- least squares
- matrix decomposition
- nonnegative matrix
- spectral clustering
- principal component analysis
- data sets
- data matrix
- document clustering
- low rank approximation
- original data
- sparse coding
- objective function
- sparse representation
- semi supervised
- optical flow
- multiscale
- computer vision
- information retrieval