Identifying Information Spreaders in Twitter Follower Networks

A number of research efforts on Twitter have been contributed towards understanding various factors that are related to retweetability, analyzing retweeting and diffusion patterns, predicting retweets, etc. One fundamental research question remains untackled: given a user and her followers, which of the followers are likely to spread her tweets to the world (the information spreader identification problem)? Answering this new and open problem helps to bridge the gap between analyzing retweetbility and understanding information diffusion. Using a large scale Twitter data set, we first find that retweet history is not an ideal method for identifying information spreaders, especially for the long tail users. Backed by statistical analysis, we set forward to extract meaningful features and present a set of feasible approaches for identifying information spreaders in the Twitter follower networks. Our study reports interesting findings, sheds light on many practical applications, helps understand the mechanisms of relaying information from one user to her followers, and offers future lines of research.