Fake News Detection an Effective Content-Based Approach Using Machine Learning Ensemble Techniques

Abstract

Any information that has been produced with the intention of deceiving readers and disseminating a concept for the purpose of gaining monetary advantage (usually political or financial) is considered fake news. Knowledge acquisition and dissemination are quick and almost free in today's society. More people than ever are using the internet. Internet channels are therefore ideal for spreading knowledge to a larger audience. What was once restricted to a select few may now instantly spread across the globe. This benefit came at the same time that its users started using false news more frequently, which is not good for a healthy society. Consequently, excellent algorithms are required to immediately identify and remove fake content. This study plans to track down an answer for the issue via mechanizing the most common way of distinguishing counterfeit news in view of its substance. Measures like the f1-score, exact grouping exactness, accuracy, and review are utilized to assess the methodology's presentation. With an accuracy of 96.7 percent, precision of 96.2 percent, recall of 97.5 percent, and f1 score of 96.9 percent, the machine learning(ML) technique played out the best in the ISOT dataset.