Identifying user behavior in online social networks

Abstract

Online social networks pose an interesting problem: how to best characterize the different classes of user behavior. Traditionally, user behavior characterization methods, based on user individual features, are not appropriate for online networking sites. In these environments, users interact with the site and with other users through a series of multiple interfaces that let them to upload and view content, choose friends, rank favorite content, subscribe to users and do many other interactions. Different interaction patterns can be observed for different groups of users. In this paper, we propose a methodology for characterizing and identifying user behaviors in online social networks. First, we crawled data from YouTube and used a clustering algorithm to group users that share similar behavioral pattern. Next, we have shown that attributes that stem from the user social interactions, in contrast to attributes relative to each individual user, are good discriminators and allow the identification of relevant user behaviors. Finally, we present and discuss experimental results of the use of proposed methodology. A set of useful profiles, derived from the analysis of the YouTube sample is presented. The identification of different classes of user behavior has the potential to improve, for instance, recommendation systems for advertisements in online social networks.