Affiliations: University of Toulouse, IRIT (Institut de Recherche en Informatique de Toulouse), SIG Team (Generalized Information Systems), 118, Route de Narbonne, F-31062 Toulouse, France, E-mail: email@example.com, firstname.lastname@example.org, email@example.com, firstname.lastname@example.org
Abstract: Existing works about user profile modeling always model the user as an independent entity. However, in social sciences, many works show the user's behavior as strongly influenced by his social ties and/or social interactions. These results were difficult to integrate in user profile modeling because of the problem of capturing social data about users in infor-mation systems. With the advent of the social web for example, this barrier can be broken on the Web, and we can reasonably think about enriching users' profiles from social data. In this paper we propose a technique to develop users' profiles with social data. We built profiles from textual users' activities data, and we show through visualization of temporal graphs the relevance of social ties on built profiles (with an experimentation carried out on 7, 081 Facebook profiles). These results motivated the potential implementation of new profiling techniques based on users' social networks, to enrich users' profiles and solve pending problems such as ‘cold start problem’ in personalized or recommender systems.
Keywords: User profile, social network, social web, Facebook, graphs visualization, dynamic graphs