Abstract: Low-credibility gossip information in social networks is a trigger for online opinion distortion. Since fragmented information always appears along with low content credibility and gossip, this paper put forward an online opinion distortion model based on DW algorithm, and information credibility, gossip intensity and opinion leaders were integrated in order to find out the mainly factors which induce opinion distortion. Then, four groups of experiment by using MATLAB were conducted, and the results show that the credibility of information is a key factor affects the speed of opinion, and also very important for the direction of the opinion dynamic of the whole network. Online opinions can easily impact by the gossip information, and extreme gossip information causes serious division of opinion group, and the whole network opinions cannot be converged even in a quite long-time iteration. Super opinion leaders have the ability to balance the influence of gossip, even in an online environment full of gossip information. Based on these, a case study including 315 thousand records of posts and comments from Weibo is conducted for confirming the mainly findings in the simulation analysis, and strategies like high credibility information, low gossip probability, and behalf opinion leader are put forwarded finally for reduce opinion distortion.
Keywords: Opinion distortion, social network, low-credibility, gossip, simulation analysis
