Developing Machine Learning Model for Predicting Social Media Induced Fake News

Fake news has been associated with major global events such as Covid-19 and the political polarisation of the US presidential election in 2016. This paper investigates how fake news has affected society and advance understanding of the nature of its impact in the future of democratic societies. Taken from large datasets consisting of over 23,000 fake news story words and over 21,000 true news story words we use descriptive and predictive analytics, partly analysing more than 350 words during the selected period of October 2016 to April 2017. The findings show that Trump was the most popular word for both true and fake news. In this study, we compare and contrast the words used and the volume of true versus fake news stories related to the election and the inauguration. This study makes an important contribution as it develops a predictive model that highlights the severity of political polarization and its consequences in democratic societies, which inevitably have implications for inclusive societies in the 21st century. © 2021, IFIP International Federation for Information Processing.