Generating social network graphs

In this paper we present and evaluate a social network model which exploits fundamental results coming from the social anthropology literature. Specifically, our model focuses on ego networks, i.e., the set of active social relationships for a given individual. The model is based on a function that correlates the level of emotional closeness of a social relationship to the time invested in it. The size of the social network is limited by the time budget a person invests in socializing. We exploit the model to define a constructive algorithm to generate synthetic social networks. Experimental results show that our model satisfies, on average, known properties of ego networks such as the size, the composition and the hierarchical structure. We also introduced a procedure for the integration of different ego networks and the generation extended social networks.