This GitHub repository hosts two novel implementations for the construction of hierarchical topic models (HTMs), HTM with word selection (HTM-WS) and HTM with document selection (HTM-DS), which can be used to build hierarchical extensions of existing topic modeling algorithms.
Both methods follow a three-step process:
- Construction of a level-1 topic model with any existing topic modeling algorithm and selection of a topic for expansion.
- Construction of the synthetic training corpus for the level-2 topic model according to HTM-WS or HTM-DS principles.
- Training of the synthetic corpus to generate the level-2 model.
At the time being, this repository provides implementations for:
| Name | Source code |
|---|---|
CTMs (Bianchi et al. 2020, 2021) |
https://github.com/MilaNLProc/contextualized-topic-models |
LDA (Blei 2003) |
https://mimno.github.io/Mallet/topics.html |