A Review of Fake News Detection Using Machine Learning Techniques

With the growing number of users of the internet people share millions of posts, articles and videos. These posts are shred on a number of social media platforms along with twitter, facebook, youtube and other social networking web sites. It is now a well-known fact that many times any misinformation may even lead to conflicts and it also has a significant influence on public opinion. The propagation of fake news stories on social media platforms and on Internet is duping people to an extent, stopping which is the need of the hour. The research area of fake news detection is gaining interest but at the same time it involves a number of challenges due to unavailability of quality resources such as datasets, published literature etc. The existing systems are not that much efficient in the detection of fake news because of the lack of fake news datasets that are comprehensive and at the same time are community-driven datasets. This has become one of the major roadblocks in the research works related to fake news detection. At the same time there are some restrictions on the input and the news category that makes it less varied. The fake news detection system aims to use data repositories such as Buzzfeed, Politifact, CREDBANK, FakeNewsNet and various classification techniques such as Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Decision tree (DT), Logistic Regression (LR), RandomForests etc. to help to achieve maximum accuracy. This review paper provides a detailed review of various fake news detection techniques used by different researchers, the datasets they have worked upon and various evaluation parameters used by them for performance evaluation of their models. We have also discussed the difficulties and challenges faced in fake news detection.