A Model for Evaluating Fake News

Abstract

"Fake news” (FN) is slowly being recognized as a security problem that involves multiple academic disciplines; therefore, solving the problem of FN will rely on a cross-discipline approach where behavioral science, linguistics, computer science, mathematics, statistics, and cybersecurity work in concert to rapidly measure and evaluate the level of truth in any article. The proposed model relies on computational linguistics (CL) to identify characteristics between “true news” and FN so that true news content can be quantitatively characterized. Additionally, the pattern spread (PS) of true news differs from FN since FN relies, in part, on bots and trolls to saturate the news space. Finally, provenance will be addressed, not in the traditional way that examines the various sources, but in terms of the historical evaluations of author and publication CL and PS.