Fake News Detection on Twitter

In this modern era, misleading news is easily transmitted and spread through different social media platforms and has become a big threat to the reliability of the news. Because of the growth of online social media, fake information does have a significant impact on society. Twitter is one of the social media platforms which is often used to propagate fake information during election campaigning. This work suggests an approach that can detect fake news by providing the probabilities for the best decision that has to be made while deciding whether an article is fake or not. The proposed work exploits not only the textual features (such as writing style and emotions) of tweets but also the characteristics of users (such as followers count and verified profile) who propagate fake news. Various computational techniques, including long short-term memory (LSTM), hierarchical attention networks and natural language process (NLP), are utilized to design the fake news detection system with improved accuracy. Finally, in order to evaluate the effectiveness of the proposed method, we have extracted about 100,000 tweets related to Haryana Assembly Elections (from 1st October to 23 October 2019) with different hashtags using Twitter’s developer API. Empirically obtained results legitimize the validity of the proposed method.