Who Are Influentials on Micro-Blogging Services: Evidence from Social Network Analysis

User influence measurement on Micro-blogging can be classified into two types according to their concerns on following relationships or interactions among users. The former focuses on counting the audiences subscribe to a user’s tweets, neglecting the fact that many audiences do not read the tweets they subscribed to, while the latter emphasizes the audiences’ explicit actions such as retweeting or replying, excluding those who read the tweets but choose to be lurkers. Since user interactions indicate that audiences have read the tweets explicitly, and following network is used to estimate the audiences who read the tweets but implementing no explicit actions, to combine the advantages of these methods, an information diffusion approach is proposed to measure user influence in this paper. Simulation results demonstrate that our proposed approach can estimate the effective audiences in the process of tweet diffusion. In the view of information diffusion, the hybrid method gives a more accurate user influence measurement by integrating following network and user interaction.