Smooth Nonnegative Matrix Factorization for Unsupervised Audiovisual Document Structuring.
Slim EssidCédric FévottePublished in: IEEE Trans. Multim. (2013)
Keyphrases
- nonnegative matrix factorization
- document clustering
- probabilistic latent semantic indexing
- negative matrix factorization
- data representation
- matrix factorization
- least squares
- semi supervised
- information retrieval
- nonnegative matrix
- principal component analysis
- information retrieval systems
- text documents
- text mining
- data sets
- unsupervised learning
- supervised learning
- spectral clustering
- clustering method
- signal processing
- image representation
- collaborative filtering
- natural language
- support vector
- image segmentation
- machine learning