Early Detection of Fake News "Before It Flies High"

Abstract

Currently, social media for news consumption is preferred over the conventional media and attracted many people due to its low cost, easy access, simplistic way of commenting & sharing, more timely nature, and rapid information sharing capabilities. On the other hand, it aggravates the prompt and wide spreading of fake news. Fake news may be fabricated for the purpose of, commercial gain, political propaganda, seeking attention, and intent of defamation. Interest of individuals, and various groups to influence events and policies around the globe is the other reason for fake news generation and dissemination. The extensive spread of fake news is progressively becoming a threat to individuals and society as a whole. It disrupts the authenticity balance of the news ecosystem; induces biased or false beliefs into consumers; creates real-life fears in the society and threatens freedom of speech, freedom of the press and democracy. The craving to mitigate the undesirable effects of fake news, recently makes fake news detection on social media an emerging research area attracting tremendous attention. Following this warm concern, various researches have been conducted and showed promising results. In this work, we propose a model for early detection of fake news using deep learning, and news content. Deep learning and heterogeneous dataset has been used to create a more generic model that could perform better in the real world. We conducted experiments on two real world datasets and a third dataset which is obtained by combining the two datasets and randomly shuffled them. Our experiment results have shown that early detection of fake news using news content and deep learning models, without waiting for news propagation, is achievable and should be given better attention to combat fake news effectively before it proliferates and misleads many people. The experimental results obtained are interesting.