User influence analysis on micro blog

In this paper, we investigate features and propose a method to identify influential users on Sina-Weibo, one of the most famous micro-blogging services in China. We first investigate features such as users' follower number distribution, relation between Weibo number and follower number and analysis of user interaction. Due to the existing methods are not very comprehensive in measuring the influence of user, we propose a new model. In which, we take the three basic actions: following, retweeting and commenting into consideration. Based on the weight and networks of them, we construct a weighted network, then employ Weighted PageRank and Hypertext Induced Topic Selection algorithm to calculate user influence. Compared with other two methods, the experiment results suggest that our model offers a new way to identify influential user, and it is more comprehensive and stable than the other two.