Emotion-Aware Fake News Detection on Social Media with BERT Embeddings

Fake news dissemination on social media poses a significant threat to the integrity of information and public discourse. This research proposes an emotion-aware fake news detection model using BERT embeddings. Leveraging the power of BERT, our model captures contextual relations in text, enabling accurate classification of fake news. Through experimentation with different BERT models, “bert-large-cased” emerges as the top-performing variant, achieving a remarkable training accuracy of 98% and an F1 score of 0.77. Integrating emotion-aware features enhances the model's efficacy in identifying fake news while minimizing false positives and negatives. Our study contributes to the field of fake news detection, offering a potent tool for safeguarding social media from disinformation.