Alessandro Vezzani — IMEM CNR Parma#
Large fluctuations in the evolution of time varying social networks
#
The complex evolution of social networks is affected by two non-trivial mechanisms: the reinforcement of ties (people tend to distribute their social events toward already contacted nodes) and the burstiness of interactions (the inter-event times are broadly distributed). Here we define and analyze an activity-driven model, featuring both characteristics. We build the model and analytically solve the dynamics in the limit of large time and connectivity. We find a non-trivial phase diagram due to the interplay of the two effects and we determine the leading contribution depending on the system parameters. Interestingly, if ties reinforcement is sufficiently strong, burstiness can be suppressed and it can play a sub-leading role also in the presence of large fluctuations. We test our results against numerical simulations, and we compare the analytical predictions with an empirical datasets (Twitter mentions network) finding a good agreement between the two.