A clustering-based analysis of a social internetworking scenario

Abstract

Social network analysis is an emergent research stream aiming at studying the characteristics of social networks. Several methodologies have been developed to extract useful knowledge from single social networks, but little attention has been devoted to studying multi-social-network scenarios, whose importance is, by contrast, dramatically growing, because social internetworking systems may offer progressively more powerful and innovative services. In this scenario, we consider clustering as a fruitful point of view from which multi-social-network scenarios can be studied. Indeed, every social network includes distinguishing features and a clustering-based approach may highlight the differences among social networks, yet analysing information about the whole system. But when we apply clustering to a social internetworking scenario (SIS), several issues arise. In this paper, we address this problem, giving interesting hints on how clustering-based analysis should be conducted and analyse a real-life SIS obtaining meaningful and original results.