A simple python implementation of LDA model with gibbs sampling and variational EM inference algorithms. The scripts have been tested on a small dataset, but it will take a long time to run on larger datasets since there aren't any computation optimization techniques.
To just verify the inference algorithm:
python lda_gibbs_sampling.py
python lda_variational_inference.py
Then the topic-word distribution will be exported as csv
file.