Fake News Detection and Analysis using Multitask Learning with BiLSTM CapsNet model

In today's world, information is of paramount importance in any field and controls our lives. With increasing number of people taking to social media and websites hosted on the internet as their everyday sources of information, the impact of spreading misinformation has high affecting causes rapidly. The proposed system is a Multitask Learning model that can categorize the news articles collected from the web as fake or not. The title and content of the articles are modeled as BiLSTM subtasks and the CapsNet model is the meta classifier. The model predicts with an accuracy of 97.96% which helps to flag the articles posted on the internet so that the readers are well informed. Further, the system is also able to rate the articles on a 5-point scale to determine the degree of misinformation in the article for further analysis.