DETECTING FAKE NEWS AND FAKE REVIEWS THROUGH LINGUISTIC STYLES

Abstract

Deceptive content has become challenging to deal with in recent years. Fake reviews continue to misguide customers on the credibility of the product. Since such data can be easily generated and is usually in abundance, fake reviews or the opinion spam problem has now become a growing research area. Also, 2016 US presidential elections proved that fake news can have a huge impact and drew attention of people to this problem. There is a pressing need for fake news detection but it is a challenging problem as well. In this paper, machine learning based classifiers have been used to automatically detect fake content (mainly fake news and fake reviews). 55 features have been extracted from data and 6 classifiers have been used for three datasets. Datasets used are publicly available and they are for fake reviews as well as fake news.