Minimum description length as an objective function for non-negative matrix factorization.
Steven SquiresAdam Prügel-BennettMahesan NiranjanPublished in: CoRR (2019)
Keyphrases
- negative matrix factorization
- minimum description length
- objective function
- nonnegative matrix factorization
- information theoretic
- mdl principle
- principal component analysis
- sparse representation
- matrix factorization
- document clustering
- tensor factorization
- gradient method
- spectral clustering
- reconstruction error
- constrained least squares
- data mining
- mutual information
- latent semantic space
- signal processing
- collaborative filtering
- feature vectors
- pattern recognition
- feature extraction