Inference, study and prediction of the dynamics of the Twitter mention network
The code available in this repository is at the base of the work that has been published in the following paper:
H. Hours, E. Fleury, M. Karsai, Link prediction in the Twitter mention network: impacts of local structure and similarity of interest 16th IEEE International Conference on Data Mining (ICDM) - DMHAA Workshop, Barcelona 12-15 December (2016).
In particular, one can find the different modules for
- exctracting mentions from Tweets (json file)
- Building mention newtork from mentions
- Inferring hashtag categories from tweets
- Computing hashtag similarity (homophily) and link overlap
- Predicting the evolution of the mention network
- Code for generating the different plots presented in the paper