The Influence in Twitter: Are They Really Influenced?

Abstract

Twitter is a popular social network service which is continuously growing. Because Twitter has become an efficient platform for advertising companies as a new vast medium, it is obvious that finding influential Twitter users and measuring their influence are important. Intuitively, users who have more followers are likely to be more influential. However, the number of followers does not necessarily mean the confidence of influence. In order to find influential users in Twitter more precisely, in this paper, we present an improvement of PageRank, which we call InterRank Inter action Rank . It considers not only the follower relationship of the network but also topical similarity between users from tweet context. By using retweet information, we verify that topical similarity indeed affects the influence of a user. Then, we compare InterRank to PageRank with an assumption that influential users are more interactive with their followers. Our comparison results show that the users found by InterRank are more interactive than those by PageRank. Overall, we believe InterRank can be an attractive alternative of PageRank in finding influential users.