'HOAXIMETER'-An Effective Framework for Fake News Detection on theWorld WideWeb

Fake news and misinformation have become a serious predicament, especially, in recent times as it has become easier to spread and harder to recognize. Not only has it created an environment of distrust around the world and misled people, but also incited violence and resulted in people losing their lives. State of the art fake news detection includes the use of verifying news from a trustable dataset, BERT filtering or using cues from Lexical Structure, Simplicity and Emotion and more. In addition, several frameworks have also been proposed to deal with this issue e.g., SpotFake and FR-detect to name a few. However, these frameworks strongly rely on the users to verify the news, modification of information, categorization of the news, and reliability of the dataset (if a dataset is used). This paper proposes 'HOAXIMETER', a framework to detect fake news on the World Wide Web covering the weakness of the aforementioned existing ones. It does so by putting forward an in-depth study and literature review on the existing frameworks by analyzing and evaluating them in the context of how well they work with each step, thereby highlighting their strengths and weaknesses. Ultimately, `HOAXIMETER' is proposed which is meant to be the most effective fake news detection framework free from the issues of the existing one.