Non-negative matrix factorization extended by sparse code shrinkage and weight sparsification non-negative matrix factorization algorithms.
Botond SzatmáryBarnabás PóczosJulian EggertEdgar KörnerAndrás LörinczPublished in: ECAI (2002)
Keyphrases
- negative matrix factorization
- nonnegative matrix factorization
- sparse representation
- sparsity constraints
- tensor factorization
- matrix factorization
- principal component analysis
- document clustering
- update rule
- latent semantic space
- blind source separation
- spectral clustering
- data clustering
- dimensionality reduction
- constrained least squares
- information retrieval
- image classification
- least squares
- object recognition
- pattern recognition