This project implements Gibbs sampling inference to LDA(Latent Dirichlet Allocation).
To-do:
- Chenk convergence
- speed up Gibbs sampling process
Reference:
@article{heinrich2005parameter, title={Parameter estimation for text analysis}, author={Heinrich, G.}, journal={Web: http://www.arbylon.net/publications/text-est.pdf}, year={2005} }
Note:
- The Gibbs sampling is very slow and it is hard to check convergence.
- The result is not very good; maybe because the corpus is not very large.
- The result can be very different in different runs.
Topic modeling tools:
- David Blei's collection: http://www.cs.princeton.edu/~blei/topicmodeling.html
- Mallet from UMass: http://mallet.cs.umass.edu/
- Stanford Topic Modeling Toolbox: http://nlp.stanford.edu/software/tmt/tmt-0.4/
- Matlab Topic Modeling Toolbox by Mark Steyvers and Tom Griffiths: http://psiexp.ss.uci.edu/research/programs_data/toolbox.htm
- LDA-J : http://www.arbylon.net/projects/
- R package: topicmodels and Topic models in R
- topic-modeling-tool(A grapical user interface tool based on Mallet): http://code.google.com/p/topic-modeling-tool/