Rumor Detection on Social Media: A Multi-view Model Using Self-attention Mechanism

With the unprecedented prevalence of social media, rumor detection has become increasingly important since it can prevent misinformation from spreading in public. Traditional approaches extract features from the source tweet, the replies, the user profiles as well as the propagation path of a rumor event. However, these approaches do not take the sentiment view of the users into account. The conflicting affirmative or denial stances of users can provide crucial clues for rumor detection. Besides, the existing work attaches the same importance to all the words in the source tweet, but actually, these words are not equally informative. To address these problems, we propose a simple but effective multi-view deep learning model that is supposed to excavate stances of users and assign weights for different words. Experimental results on a social-media based dataset reveal that the multi-view model we proposed is useful, and achieves the state-of-the-art performance measuring the accuracy of automatic rumor detection. Our three-view model achieves 95.6% accuracy and our four-view model using BERT as a view also reaches an improvement of detection accuracy.