Anempirical Analysis of Classification Models for Detection of Fake News Articles

Fake news has been at the centre of the debate raging on how humans consume information in the digital era. The rise in use of social media has made it easier to blur the line between what is and what isn't a reliable source of information, making this one of the most pressing issues of our time. This paper approaches the issue from a data-oriented perspective by investigating whether automatic computational approaches in NLP and Machine Learning can be used to detect falsehoods in written text. Performance of features like n-grams and word vectors used with five supervised learning techniques in detecting Fake News articles are compared. The impact of certain changes in the parameters of feature extraction on classifier performance are also analysed in this paper.