Non-negative matrix factorization algorithms greatly improve topic model fits.
Peter CarbonettoAbhishek SarkarZihao WangMatthew StephensPublished in: CoRR (2021)
Keyphrases
- topic models
- negative matrix factorization
- latent dirichlet allocation
- topic modeling
- nonnegative matrix factorization
- learning algorithm
- computer vision
- prior knowledge
- probabilistic latent semantic analysis
- pattern recognition
- probabilistic model
- information extraction
- data sets
- artificial intelligence
- spectral clustering