Microblog, as one of the most popular social networking services, plays an increasingly significant role in communication and information propagation. Among the numerous studies on social network, one critical problem is identifying the influencers (or opinion leaders) and quantifying the influence strength of each individual effectively. This paper focuses on the problem of measuring users' influence in microblog network. We define respectively, the social influence in terms of the ability of interactivity driven and the breadth of information dissemination in global network. With the analysis of the characteristics of interactive behaviors and the way of information spread, the principal factors indicating influence are explored, which include the quantity and quality of followers, the quality of tweets, the ratio of retweeting and the similarity of users' interests. Although so many metrics have been taken into account to measure influence in proposed User Relative Influence Measure Model and User Network Global Influence Model, our Influence Rank Algorithm to implement the models is only O(e) on time complexity. Finally, the experimental evaluations and comparisons with related algorithms on million-user-level dataset demonstrate the efficiency and effectiveness of Influence Rank Algorithm.
展开▼
