Identifying and Classifying Fake COVID-19 Tweets using Transformer Models

COVID-19 pandemic has significantly impacted daily life around the world during the past few years. According to government measures, people used to prevent spreading the virus by quarantine, social distancing, staying at home and distance learning. The problem of fake news in social media is dramatically increasing during this pandemic, especially when it refers to fake news about Covid-19. The objective of this research is to propose a transformer model to classify and identify fake covid-19 tweets. BERT, RoBERTa, and DistilBERT transformer models are applied. The results showed that RoBERTa outperforms the other two transformer models by achieving 97% accuracy and ROC AUC Score of 0.961.