Contrastive topic evolution discovery via nonnegative matrix factorization.
Qianru LiBo ChenSongjun MaLuoyi FuXinbing WangPublished in: ICC (2016)
Keyphrases
- nonnegative matrix factorization
- probabilistic latent semantic indexing
- matrix factorization
- negative matrix factorization
- data representation
- nonnegative matrix
- active set
- least squares
- principal component analysis
- topic models
- spectral clustering
- document clustering
- sparsity constraints
- original data
- latent dirichlet allocation
- knowledge discovery
- objective function
- co occurrence
- low rank
- information extraction
- factor matrices
- data mining