Applying non-negative matrix factorization on time-frequency reassignment spectra for missing data mask estimation.
Maarten Van SegbroeckHugo Van hammePublished in: INTERSPEECH (2009)
Keyphrases
- missing data
- negative matrix factorization
- matrix factorization
- principal component analysis
- low rank
- nonnegative matrix factorization
- missing values
- structure from motion
- incomplete data
- sparse representation
- motion segmentation
- gradient method
- multiple imputation
- document clustering
- stochastic gradient descent
- tensor factorization
- spectral clustering
- data matrix
- database
- model selection
- signal processing
- natural language processing
- pairwise
- matrix completion
- keywords