Identifying Fake News On Social Media – A Comparative Analysis Using Machine Learning

Authors

  • Saransh Bansal

Abstract

Online networking for news utilization is a double-edged sword. Firstly, its lack of
trustworthiness, direct access, and high-speed lead to the spread of misinformation through
interpersonal communication. Secondly, data spread through the online network is noisy, lacking,
unstructured and large. The expansive spread of false news has the potential of detrimental
consequences for individuals, society and business. In this context, we show a far-reaching approach
for identifying fake news by conducting a comparative analysis of the performance of various
classifiers such as Gradient Boosting Classifier, Random Forest Classifier, Logistic Regression,
Passive Aggressive Classifier, Support Vector Machine (SVM) and Stochastic Gradient Descent
using SVM methodology on the highly reliable publicly available Kaggle dataset. Based on the
selected features of text classification and machine learning, we recommend SGD-SVM model
which can be utilized by organizations, especially social media marketing firms to detect fake news.

Published

2020-11-01

Issue

Section

Articles