In this seminar I will show how local and global properties of networks can be obtained as a result of particular social characteristics of the nodes. In particular I will present three
different network growth models able to reproduce some typical topological properties observed in real networks starting from the individuals? choices.
A typical characteristic of real social networks is the presence of a power law degree distribution with a marked exponential tail. The classical Barabasi-Albert model can reproduce the power law structure.
The exponential tail is usually interpreted as a finite size effect in this procedure. I will show that, on the contrary, this typical structure, can be originated by as an effect of the limited knowledge, of the entering individuals of the pre-existing nodes in the network. Several network structures have an important dependency from geography, for example commuting networks. I will show how the social preferences of the individuals are able to reproduce the spatial properties of the networks.
Finally I will show a network generation method able to reproduce local homophily properties in a social contact graph.


Floriana Gargiulo

Invited Talk e-session


Photos by :