Studying the Global Spreading Influence and Local Connections of Users in Online Social Networks

In online social networks (OSNs), highly-connected users are generally more capable to trigger viral diffusion. However, recent research demonstrates that ordinary users who only have a few connections can also cause large-scale diffusion. In this paper, we study the relation between the global spreading influence and the local connections of users to theoretically explain this phenomenon. We focus on two fundamental but important measures of local connections: degree and assortativity. Degree counts the number of connections of individual users, assortativity measures the connection patterns between users. To measure the global spreading influence of users, we adopt k-shell which assigns a coreness ks to each user. We find that, coreness ks shows a power-law dependence on the average degree ‹dk› of users on shell ks: ks ∞ ‹dk›β, with 0.5