Real or Fake: An intrinsic analysis using supervised machine learning algorithms

Abstract

Different platforms of information data semantics are affected by Fake news in the recent years. Due to the inherent writing style and propagation speed of such false information, it has been difficult to pinpoint them from the true ones. The related work in the field makes use of various supervised as well as unsupervised machine-learning algorithms to classify and detect fake news. This paper provides an in-depth overview of the algorithms that are being used for detection. The paper also provides an analysis of notable algorithms on two datasets: Source based Fake News classification and Fake and Real News dataset. The results show that supervised algorithms with proper embedding and vectorizer models can provide great accuracies. The experimentation output shows the effectiveness of the proposed architecture.