Discovering Shared Interests in Online Social Networks

The capacity of rapidly disseminating information such as latest news headlines has made online social networks a popular and disruptive venue for spreading influence and distributing contents. Given the importance of online social networks, it becomes increasingly imperative to understand the shared interests of users on the popular information or contents that circulate through these networks. This paper proposes a novel graphical approach based on bipartite graphs and one-mode projection graphs to model the interactions of users and information and to capture the shared interests of users on the information. The experiments based on data-sets collected from Digg, a popular social news aggregation site, have demonstrated the proposed approach is able to discover inherent clusters of users and information within online social networks. The evaluation results also show that these clusters exhibit distinct characteristics. To the best of our knowledge, this paper is the first attempt to apply bipartite graphs and one-mode projections to shed light on the interactions of people and information in online social networks and to discover the clustered nature of users and contents.