Information vs Interaction: An Alternative User Ranking Model for Social Networks

The recent years have seen an unprecedented boom of social network services, such as Twitter, which boasts over 200 million users. In such big social platforms, the influential users are ideal targets for viral marketing to potentially reach an audience of maximal size. Most proposed algorithms use the linkage structure of the underlying network to measure the information flow and hence evaluate a users influence. Yet that is not the full story for social networks. In this paper, we propose to examine users' influence from a social interaction perspective. We built a ranking model based on the dynamic user interactions taking place on top of these underlying linkage structures. In particular, in the Twitter setting we supposed a principle of balanced retweet reciprocity, and then formulated it to re-evaluate the value of Twitter users. Our experiments on real Twitter data demonstrated that our proposed model presents different yet equally insightful user ranking results.