Combating Fake News with Machine Learning and Deep Learning Methods

摘要: Of late, there is incessant nuisance in the public domain due to spreading of fake news. Misinformation in the Political and Health sectors can create a lot of havoc. Social media is making this spread more rampant, and there are not many practical means to put a check on the same. In this situation, combating fake news has become a necessity. We explore the route of applying machine learning and deep learning algorithms to identify fake news from real news. From our experiments, we find that LSTM algorithm is able to give an accuracy of 61%. We conclude that when the dataset used for model building is properly labeled, and the appropriate algorithms are used for training, then identification of fake news is possible. Based on this static analysis, we can further improve to develop a dynamic/online model that can help in making the detection more real-time.