Misinformation Detection Using an Ensemble Method with Emphasis on Sentiment and Emotional Analyses

Abstract

As of April 2023, over 6 million people have lost their lives due to COVID-19 according to the World Health Organization (WHO). With no prior knowledge of this disease, people have turned to the Internet including social media to search for available remedies. However, it is important to note that the Internet cannot replace primary healthcare providers as there is a significant amount of false information. This research proposes a system to identify fake news by combining the results from several ensemble learning methods (including bagging, boosting, stacking, & voting means) and recurrent neural network (RNN). Additionally, sentiment and emotional analyses are employed to determine whether the accuracy of fake news detection can be improved. Experiment results show the ensemble learning methods provide higher accuracy than standalone RNN model. Moreover, this study reveals that incorporating sentiment and emotional analyses in fake news detection improves the accuracy of misinformation identification.