A BERT Classifier Approach for Evaluation of Fake News Dissemination

The intake of information via media has changed from newspapers to social networking sites as technology has evolved. Accessibility and availability are two important aspects that have led to this trend in media usage. Users share thousands of posts, articles, and videos as the internet penetration grows. These postings are done on a variety of social networking sites such as Facebook, Instagram, YouTube, Twitter, and others. It is now widely acknowledged that disinformation may often create tensions and has a substantial impact in the country. Stemming the tide of false news articles via social networking sites and the Internet is critical. This problem has indeed been handled in this study using various algorithmic approaches that could help us to control this type of hazardous thing. Including the findings, a contrast of how various classification functions are offered. This article gives a thorough assessment of numerous fake news detection strategies employed by various other authors, databases they had engaged with, and the multiple analytical people were using to evaluate the effectiveness of their respective algorithms. This research analyses the troubles and challenges of detecting this type of news. This article examines documents from 2017 through 2022, as well as various fake news detecting tools. This study provides a complete evaluation of present and past studies on false news recognition leveraging various ML and DL models. In this research work, BERT Classifier has been used which uses a deep learning technique and has the highest accuracy of all the methods described.