HiTR: Hierarchical Topic Model Re-estimation for Measuring Topical Diversity of Documents.
Hosein AzarbonyadMostafa DehghaniTom KenterMaarten MarxJaap KampsMaarten de RijkePublished in: CoRR (2018)
Keyphrases
- topic models
- topic modeling
- text documents
- latent topics
- probabilistic topic models
- latent dirichlet allocation
- pitman yor process
- topic discovery
- text mining
- document classification
- lda model
- probabilistic model
- relevance model
- generative model
- latent variables
- author topic model
- text corpora
- co occurrence
- statistical topic models
- text analysis
- information retrieval
- latent semantic analysis
- text collections
- news articles
- word pairs
- document clustering
- machine learning
- collaborative filtering
- variational inference
- information retrieval systems
- generative process
- microblog posts
- gibbs sampling
- document collections
- text classification
- text streams
- dimensionality reduction
- baseline models
- keywords
- concept hierarchy