Documentation, usage examples and certain features are work in progress. The code isn't in great shape, the aim of this early release is to get ideas on potential applications and feedback on what features would be useful to users.
Initial demo here.
- Extract hierarchies of topics from corpuses of plaintext news articles.
- Visualize trends in topics over time.
- Automatically tag new unseen articles with topics.
- Determine "interesting" topics in a corpus.