Comparative analysis of machine learning algorithms to detect fake news

Abstract

Fake news has immense impact in our modern society. The widespread dissemination of false news has the potential to have highly damaging consequences for both individuals and society. As the readers come across many fake news when they come across a real news, they believe that it could be another fake news. The aim of this project is to perform a comparative analysis of three algorithms (Multinomial Naive Bayes, Passive Aggressive Classifier and Decision Tree Classifier) using Natural Language Processing techniques to develop a solution that users can use to identify false or misleading information. As Passive Aggressive Classifier gave the best results, prediction is done using this classifier.