This repository contains the code used the run the experiments in Bayesian Nonparametrics for Sparse Dynamic Neworks
In order to run the reddit and reuters experiments, download the soc-redditHyperlinks-title.tsv file from https://snap.stanford.edu/data/soc-RedditHyperlinks.html and the Days.net file from http://vlado.fmf.uni-lj.si/pub/networks/data/CRA/terror.htm respectively, and place them in the data folder.
Then, run load_reddit.py or load_reuters.py to preprocess the data.
Finally, run the experiments using reddit_experiment.py or reuters_experiment.py.
After running the desired number of chains, combine them using combine_chains_reddit.py or combine_chains_reuters.py. Then, the results can be plotted using the combined results using plot_results_reddit.py or plot_results_reuters.py.
The simulated data experiment can be run directly by running simulated_experiment.py, which will also plot the results.