Discriminative nonnegative matrix factorization using cross-reconstruction error for source separation.
Kisoo KwonJong Won ShinHyung Yong KimNam Soo KimPublished in: INTERSPEECH (2015)
Keyphrases
- nonnegative matrix factorization
- reconstruction error
- sparse representation
- negative matrix factorization
- objective function
- singular value decomposition
- matrix factorization
- sparse coding
- semi supervised
- principal component analysis
- least squares
- unsupervised learning
- spectral clustering
- subband
- feature extraction
- document clustering
- original data
- image classification
- high dimensional
- data sets
- dimensionality reduction
- feature selection
- image patches
- wavelet transform
- clustering algorithm
- information retrieval