FAKE NEWS DETECTION USING NLP AND OPTIMISED DEEP LEARNING ALGORITHMS

Abstract

In today’s ever-growing internet, information and news spreads rapidly like wildfire. A fraction of this news is fake or misleading, mostly for political purposes. These fake news need to be monitored. Since the amount of information generated everyday is gigantic, it is only viable to automate the fake news detection procedure. In this paper, we have thoroughly gone through various deep learning and neural network techniques that can be applicable for fake news detection. We have compared these algorithms based on there architecture and mode of operation in terms of the problem statement. Detecting fake news is an essential yet one of the most challenging task to be be done in Natural Language Processing and Machine Learning. The tremendous rise of social media platforms has enhanced the spread of fake news while also dramatically expanding the amount of information available. As a result, false news’ effect has already expanded, sometimes spilling over into media platforms and endangering public safety. Given the large volume of Web content, automatic false news detection is indeed a practical NLP task useful with any and all online copyright holders in ways that minimize work and attention in detecting and restrict the development of misinformation. The challenges of identifying fake news, and maybe even associated tasks, are discussed in this article. In this study, we offer a methodology for classifying fake news. An ensemble model comprised of pre-trained transformer based models was utilised.