Automatic Fake News Detection by Exploiting User’s Assessments on Social Networks: A Case Study of Twitter

Abstract

Nowadays, social media has been becoming the main news source for millions of people all over the world. Users easily can create and share their information on social platforms. Information on social media can spread rapidly in the community. However, the spreading of misleading information is a critical issue. There are much intentionally written to mislead the readers, that are called fake news. The fake news represents the most forms of false or unverified information. The extensive spread of fake news has negative impacts on society. Detecting and blocking early fake news is very essential to avoid the negative effect on the community. In this paper, we exploit the news content, the wisdom of crowds in the social interaction and the user’s credibility characteristics to automatically detect fake news on Twitter. First, the user profile is exploited to measure the credibility level. Second, the users’ interactions for a post such as Comment, Favorite, Retweet are collected to determine the user’s opinion and exhortation level. Finally, a Support Vector Machine (SVM) model with the Radial Basis Function (RBF) kernel is applied to determine the authenticity of the news. Experiments conducted on a Twitter dataset and demonstrated the effectiveness of the proposed method.