Detect Cooperative Hyping Among VIP Users and Spammers in Sina Weibo

Sina Weibo services provide platforms for massive information dissemination and sharing between hundreds of millions of users. The hot topics in this platform attract substantial interest and have enormous potential for business and society. As a result, it has attracted spam teams with malicious intent. In this paper, we study how to detect such opinion spam teams and how do they guide public opinion. Our model is unsupervised and adopts a Bayesian framework to distinguish between spammers and non-spammers. Experiments conducted on a dataset of a Sina Weibo hot topic with a 0.81 F1-measure demonstrate the proposed method's effectiveness. Through further analysis, we found the phenomenon and some methods of the cooperative hyping among VIP users and spammers (in Sect. 3.2). VIP users are small in number but have a great influence due to they have a large number of followers, so VIP users are responsible for hyping topics so that it attract more attention, then a lot of spammers guide public opinion through a lot of manual postings.