A DEEP LEARNING-BASED FAKE NEWS DETECTING SYSTEM

Since the birth of the Internet, social media has gradually taken up an increasingly important role in our lives. Whether it's food, clothing, housing, transportation, or keeping up with the latest events, we all rely on the vast amount of news and information provided by social media. Therefore, many unscrupulous business entities that publish all kinds of false information for profit, and the general public is easily misled because of their limited knowledge reserve. In this paper, we embrace the FNC-1 challenge as the foundation for crafting our innovative fake news detection system. In the course of our exploration, we discerned that the FNC-1 dataset was marred by issues pertaining to both class imbalance and data scarcity. To surmount these intricacies, we introduce an original data augmentation approach hinging on the principles of deep learning. Experimental results show that our proposed method outperforms state-of-the-art(SOTA) fake news detection approaches by 6.9% F1 score on the FNC-1 dataset.