Characterization, Classification and Detection of Fake News in Online Social Media Networks

Abstract

Due to its increasing popularity, low cost, and easy-to-access nature, Online Social Media (OSM) networks have evolved as a powerful platform for people to access, consume, and share news. However, this has led to the large-scale distribution of fake news, i.e., deliberate, false, or misleading information. Fake news is a pressing dilemma, as it has serious negative implications for individual users and for society as a whole. The news contents in the OSM networks are distributed rapidly, so the identification systems should predict news items as soon as possible to avoid spreading false news. Therefore, it is extremely crucial and technically challenging to detect fake news in social media networks. In this paper, we have discussed different characteristics and types of fake news and also propose an effective solution to detect fake news in OSM networks. The stance detection model and the fabricated content classifier are the main two components of the solution. The stance detection model achieved an accuracy of 90.37% with Logistic Regression, and the fabricated content classifier achieved an accuracy of 93.46% with Bi-directional LSTM.