Automatic Relevance Determination in Nonnegative Matrix Factorization with the $(\beta)$-Divergence.
Vincent Y. F. TanCédric FévottePublished in: IEEE Trans. Pattern Anal. Mach. Intell. (2013)
Keyphrases
- nonnegative matrix factorization
- automatic relevance determination
- missing data
- matrix factorization
- neural network
- discriminant analysis
- data representation
- negative matrix factorization
- feature selection
- expectation propagation
- spectral clustering
- principal component analysis
- least squares
- document clustering
- error reduction
- bayesian inference
- semi supervised
- objective function
- text mining
- natural language processing
- missing values
- original data
- classification accuracy
- image sequences
- image processing