Fake News Detection Using Social Media User Network and Engagement Features

Today, social media has become a key source of news consumption, posing a crucial issue of disseminating and propagating a large volume of fake news. As manual fact-checking techniques are slow and time-consuming, automated fake news detection models that can efficiently manage vast volumes of information must be developed. While natural language processing and computational linguistics have been widely employed in previous research, they often prove inadequate for constructing robust models. By examining the behavior of social media users, this research incorporated user network and engagement features to capture the patterns of fake news propagation on social media. We aimed to examine these features from the understanding that users are influenced by social validation and homophily, which play a significant role in shaping their behavior within online social networks. These features are then incorporated into our proposed fake news detection model. Our experimental results have demonstrated the significance of these features, suggesting their usefulness and relevant effects in fake news detection. Overall, this research contributes to the ongoing efforts in combatting fake news by proposing and assessing the significance of user network and engagement features within a fake news detection framework.