In this repository, I've run some topic modeling experiments on the NIPS papers dataset.
- python 3.x
- pandas
- nltk
- gensim
- pyLDAvis
To prepare corpus and train LDA model, first unzip data/papers.tar.xz into the data folder and then:
python topic_modeling_train.pyAfter training, you can see visualization results in lda_vis.html file.
To test model:
python topic_modeling_test.pyThere is also a jupyter notebook containing complete code for train and test process.