Fake News Detection in Social Media: A Systematic Review

Abstract

The growth of social networks platforms leverages the consumption of news due to its easy access, spreading behavior, and low cost. However, this revolution in the way that information is released has provided the growth of something that always walked side by side with the real news: we are talking about fake news. After the 2016 U.S. presidential election this term became more popular and dangerous because of its negative effect on society. In this context, recent contributions has appeared addressing several related topics, such as spreading behavior, methods for spreading contention, and fake news detection algorithms. Despite of the growth of this type of research, it is difficult for a researcher to identify the current state-of-the-art literature about fake news detection. To overcome this obstacle, this paper presents a systematic review of the literature that brings an overview of this research area and analyzes the the high-quality studies about fake news detection. Through this systematic literature review, more than 6,000 articles were found according to our search protocol. Then, we put these studies through stages of screening to ensure that they were quality assessed. Were elected 32 high-quality studies according to our PRISMA flow diagram defined in this paper. These studies were then categorized by their contribution type and algorithm. This work shown that Twitter and Weibo1 are the social media platform most applied by selected studies, and deep learning algorithms given the best detection results, specially LSTM. Besides, this SR exposes the lack of research for fake news detection in other language than english. Finally, we expect this study can help researchers identify the greatest contributions as well as research opportunities.