Fake News Detection Based on Social Features by Ordered Weighted Averaging Fusion

Today, different groups of people use social media in their businesses and normal daily activities specially for accessing news and their favorite information in various fields. Facing with huge amounts of information and news in social media makes different challenges for the users. One of the main challenges of the users is distinguishing valid news and information from invalid and fake ones. Fake news means low quality news containing inaccurate or invalid information. Because of the fast and widely spread of the news in social media, they may have very destructive effects on the user's social behavior. Therefore, the fake news should be identified and banned as soon as possible.  To overcome the challenge of identifying fake news, in this manuscript a method is introduced to use profile features of the users and some features of the tweets in twitter to determine the possibility of a tweet being fake. This method also uses ordered weighted averaging as a data fusion method to increase the accuracy of the detection. To determine the effectiveness of the presented method, some experiments are designed based on the known datasets from twitter. The evaluations of the results of these experiments indicate effectiveness of the proposed method.

Keywords:

fake news detection , Data fusion , social features , tweet features , user profile features , OWA