Building Emergent Social Networks by Semantic User Preference Clustering.pdf

Building Emergent Social Networks by Semantic User Preference Clustering Iván Cantador, Pablo Castells Escuela Politécnica Superior, Universidad Autónoma de Madrid Campus de Cantoblanco, 28049 Madrid, Spain {ivan.cantador, pablo.castells}@uam.es Abstract. This paper presents a novel approach to automatic semantic social network construction based on semantic user preference clustering. Considering a number of users, each of them with an associated ontology-based profile, we propose a strategy that clusters the concepts of the reference ontology according to user preferences of these concepts, and then determines which clusters are more appropriate to the users. The resultant user clusters can be merged into in- dividual group profiles, automatically defining a semantic social network suit- able for use in collaborative and recommendation environments. 1 Introduction The swift development, spread, and convergence of information and communication technologies and support infrastructures, reaching all aspects of businesses and homes in our everyday lives, is giving rise to new and unforeseen ways of inter-personal con- nection, communication, and collaboration. Virtual communities, computer-supported social networks, and collective interaction are indeed starting to proliferate and grow in increasingly sophisticated ways, opening new opportunities for research on social group analysis, modeling, and exploitation. In this paper we propose a novel approach towards building emerging social networks by analyzing the individual motivations and preferences of users, broken into potentially different areas of personal interest. Finding hidden links between users based on the similarity of their preferences or historic behavior is not a new idea. In fact, this is the essence of the well-known col- laborative recommender systems (e.g. see [12]). However, in typical approaches, the comparison between users is done globally, in such a way that partial, b