Identifying Important Users in Sina Microblog

Important users are high-status vertices in social networks. They are everywhere in most fields of society and have big impact on those around them. Although a lot of effort has been made on identifying important users, the efficient methods still need to be developed, especially for the web users from Sina microblog, which is the most popular social networking sites in China and has unique characteristics. In this paper, a machine learning-based method which only uses several attributes on Naive Bayes Classifiers (NBC) and Back Propagation Neural Network (BPNN) was proposed to identify important users. Initial experiments indicate that our method is effective. The result of "high" category has more than 55% accuracy rate. We find the NBC can identify more important users while BPNN has higher accuracy rate. What's more, the numbers of follower and followings in Sina microblog is independent.