Complex networks project (2019).
- think about multilayer network?
- for certain time window we can define questions
- roboustness attack: investigate how removal of few most popular answering users would collapse the network and num of questions would be unanswered
- activity driven network
- Basic analysis of undirected network (metrics, if scale-free network etc)(Almost done)(RSK)
- Define user types and try to find patterns which define user type by comparing directed and undirected graph (RSK)
- Thread detection(NCH)
- Motifs analysis (implementation + look for patterns between layers and networks)(RPY)
- Popularity prediction (VSK)
Notes:
- check for each node in undirected graph as interaction(compare directed and undirected)
- analyze motifs (algorithms from paper) and perform similar analysis as they did
- GROUND TRUTH: no of posts per year(fitting distribution to the interactions we can calculate)
- Finally analyze patterns on MathExchange and then later check if it generalize to others (super-user and ask-ubuntu)