Machine Learning Paradigm for Internet of Things Applications

10Fake News Detection Using Machine Learning Algorithms

M. Kavitha*, R. Srinivasan and R. Bhuvanya

Computer Science & Engineering, Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology, Chennai, India

Abstract

Any sorts of untrue information designed to mislead or defame any individual are termed as fake news. In this paper, we describe a technique to spot bogus news using machine learning (ML) architectures. The ever-increasing production and circulation of skewed news stories necessitates an immediate need for software that can automatically discover and detect them. Automated identification of fake news, on the other hand, is extremely difficult because it necessitates the model’s understanding of natural language nuances. Binary classification method was used as a fake news detection technique in the existing methodologies, which restricts the ability of the model to discern how closely the reported news is connected to true news. To address these issues, this paper attempted to design a neural network architecture that could reliably predict the attitude between a given combination of headlines and article bodies. We have fused some advanced ML techniques together which includes the logistic regression and recurrent neural network technique particularly. The implemented approach is better than the existing architectures by 2.5% and achieved accuracy of 90.39% on test data.

Keywords: Machine learning, binary classification, logistic regression, recurrent ...