Memory effects in time evolving networks

PLANARY SESSION

Friday 26th of September

Abstract:

In many social and information systems, the network of the interactions generated by the agents activity is a rapidly evolving time-varying structure, where memory effects and reinforced processes for strong ties drive the evolution. We extract a simple form of the memory process from extensive datasets and we propose an analytical approach to non-Markovian dynamics underlying the evolution of the network. From the master equation of the process, we obtain an analytic scaling form for the evolving degree distribution and for the average degree of the network as a function of the activity probability distribution and of the memory parameters. A striking agreement between the model predictions and both numerical and extensive real data is found.