FAKE NEWS CLASSIFIER USING MACHINE LEARNING

Abstract

Intentionally false material disguised as respectable journalism is a global information accuracy and integrity concern that influences
opinion formation, decision making, and voting habits. Social media channels like Facebook and Twitter first propagate so-called 'fake
news,' which then spreads to conventional media outlets like television and radio news. Fake news articles spread through social media often
have similar language features, such as the overuse of unsupported exaggeration and the failure to properly identify referenced information,
when they first appear. Results of a fake news detection investigation are reported in this article to demonstrate the effectiveness of a fake
news classifier. There are many tools that may be utilised to construct a new kind of false news detector that utilises quoted attribution in a
Bayesian machine learning system as one of the primary features. To put it another way, it is 63.33% accurate in detecting bogus quotations
when used in articles. Influence mining is an innovative technology that may be used to identify bogus news and even propaganda, according
to the authors. The classifier performance and findings, as well as the research procedure, technical analysis, and technical linguistics, are
all discussed in this study. After discussing how the existing system would grow into an influence mining platform, the report ends.