To run the code, it needs to have toy_dataset.text as an example and vocabulary.py for pre-processing in the same path,
You can run the code simply by:
python Coplda.py [path2train] [Number of topics] [Number of iterations] [Removing stop words] [percent of train documents]
For instance:
python Coplda.py toy_dataset.txt 20 50 false 90
Output:
1-cosine(Number of topics).txt : 1- cosine similarity between topic distributions of two consecutive documents (for example second row number is between doc0 and doc1) (near to zero means more similar)
lambda(Number of topics).txt : lambda hyperparameter between topic distributions of two consecutive documents (for example second row number is between doc0 and doc1) (the higher lambda the more similar documents)
Perplexities(Number of topics).txt (each row for each iteration)
topc10words.txt (top-10 words , each row related to one topic)
please use this bibtex to cite this work:
@inproceedings{amoualian2016kdd, author = {Amoualian, Hesam and Clausel, Marianne and Gaussier, Eric and Amini, Massih-Reza}, title = {Streaming-LDA: A Copula-based Approach to Modeling Topic Dependencies in Document Streams}, booktitle = {Proceedings of the 22nd International Conference on Knowledge Discovery and Data Mining}, series = {SIGKDD}, year = {2016}, isbn = {978-1-4503-4232-2}, location = {San Francisco, California, USA}, pages = {695--704}, numpages = {10}, url = {http://doi.acm.org/10.1145/2939672.2939781}, doi = {10.1145/2939672.2939781}, acmid = {2939781}, publisher = {ACM}, address = {New York, NY, USA}, keywords = {copulas, document streams, latent dirichlet allocation, topic dependencies}, }