A text analysis based method for obtaining credibility assessment of Chinese microblog users

Abstract

Nowadays, social network platforms, such as Weibo have become important ways for people to get information. However, considering their ability to disseminate information quickly and widely, they also become breeding ground of rumors. Thus, the information credibility on Weibo has become a meaningful issue to be paid attention to. This research investigated on the process of rumors spreading and refuting, using a new method which can quantify users’ credibility assessment. The method was based on text analysis, designing weighted sum of words frequency as microblog’s credibility score. By this method, two significant processes related to rumor spreading, the credibility dissemination process of a microblog and the change of people’s credibility assessment towards a rumor, were investigated and assessed. Shown from the results some enlightenments about refuting rumors could be concluded. The findings are: (1) the designed method is reliable in some situation. Although it is a simple, rough classification method, it can find the credibility attitude the microblog expressed and help us finding essential users. (2) Popular individuals is important in the dissemination of credibility assessment in Weibo, they can easily convey their judgement. (3) The key to rumor refutation is not only timely reaction, but also wide and sustaining spread to form consensus. If the disputation is not continuous, people may forget it soon.