https://github.com/qlai/stochasticLDA

Python implementation of Stochastic Variational Inference for LDA
https://github.com/qlai/stochasticLDA

Last synced: 3 months ago
JSON representation

Python implementation of Stochastic Variational Inference for LDA

• Host: GitHub
• URL: https://github.com/qlai/stochasticLDA
• Owner: qlai
• License: gpl-3.0
• Created: 2016-04-04T01:14:33.000Z (over 8 years ago)
• Default Branch: master
• Last Pushed: 2020-05-26T18:11:58.000Z (over 4 years ago)
• Last Synced: 2024-05-19T23:35:25.071Z (6 months ago)
• Language: Python
• Homepage:
• Size: 35.2 KB
• Stars: 25
• Watchers: 3
• Forks: 23
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.txt

Awesome Lists containing this project

README

        

##Stochastic Variational Inference for Latent Dirichlet Allocation

Code structure from the OnlineVB code provided by Matthew D. Hoffman ([email protected]) and the algorithm is as described in Hoffman's paper below

Based on the following papers:

- [Latent Dirichlet Allocation](https://www.cs.princeton.edu/~blei/papers/BleiNgJordan2003.pdf) by David M. Blei, Andrew Y. Ng and Michael I. Jordan

- [Stochastic Variational Inference](http://www.columbia.edu/~jwp2128/Papers/HoffmanBleiWangPaisley2013.pdf) by Matthew D. Hoffman, David M. Blei, Chong Wang and John Paisley

###Also aiming to implement SVI for HDP as described in the second paper above, work in progress

###How to Use

See 'Help' using

```python stochastic_lda.py -h```

You will need:

- A file [dictionary.csv] containing your vocabular

- A file [doclist.txt] containing the list of documents in the directory that you want to sample from

- At the moment your documents can be just a normal txt file, no pre-processing required

For classwork, work in progress...

- [x] Basic initial implementation

- [x] Debug for common corpus

- [x] Support Command-Line Usage for user-defined test mode and normal mode

- [x] Run on own data

- [ ] Implement HDP