Analyzing the influential people in Sina Weibo dataset

With the increasingly rapid growth of micro-blogging services, influence analysis is becoming a very important topic in this area. Sina Weibo, one of the largest mirco-blogging services in China, has provided a new operation comment-only, which allows users to give feedback on a post without forwarding. However, most of existing works focus on Twitter, which fail to consider this new operation. In paper, we propose a new influence measurement method called WeiboRank on Sina Weibo which applies comment-only operation to help researchers to find ignored influential users who have big influential in comment dimension. Furthermore we analyze why a particular user is influential based on tracing the source of influence to find out which aspects contribute to influence. Our experiments based on a subset of a whole Sina Weibo datatset, which includes three-month records of 22,514,394 users. We conduct extensive experiments and results show that we can accurately find the most influential individuals among entire social networks, while the running time of the algorithm increases linearly with any increase in data size, which is suitable for large scale networks.