PC Seminar: Friend of a friend models of network growth

Abstract: A model for a friend of a friend network growth is based on the idea that individuals joining the social network choose one individual at random and link to her with probability p, then they choose a friend of that person and link with probability q. The model is more general and conceptually simple, yet it produces power-law degree distributions, small world clustering and super-hub networks with non-stationary degree distributions. 
I will discuss a general framework for analysing a friend of a friend models of network growth and look at some special cases which produce scale-free and super-hubs networks. I will then discuss the general results of the models and show examples of misleading claims about how some cases of models similar to the friend of a friend models can be used as a form of local mechanism for motivating preferential attachment. 
Finally, I will mention some results about the early evolution of 2018/2019 Tanzania Twitter and compare it with the 2012 Twitter networks of USA, Brazil and Japan. 
The talk is based on some recent studies by Watson Levens, David J. T. Sumpter and Alex Szorkovszky. 

This is a seminar in our seminar series on Probability and Combinatorics (PC).