#Unsupervised Topic Modelling Topic modelling for scientific articles using gensim LDA, LSI and HDP implementations. Utilizes titles and abstracts of scientific articles to obtain document/topic distributions of the input data-set. Currently LDA performs the best with alpha='auto' and number of topics between 30-50.
- Gensim
- Numpy
- Sci-kit
- Plotly (For plotting graphs of topic distributions)
python topicmodeling.py <path_to_input> <path_to_saved_models_output> <number_of_topics>