Predicting the Virality of Fake News at early stage of diffusion.
(A2022-106544)

AUTHORS

Lisbeth Jimenez Rubido, Universidad Carlos III de Madrid; Mercedes Esteban-Bravo, Universidad Carlos III de Madrid; Jose Vidal-Sanz, Universidad Carlos III de Madrid

ABSTRACT

Fake news about brands and companies can cause reputation and financial damage, and a reliable early damage detection is only possible forecasting the virality of fakes news pieces, and early focusing companies’ attention on those pieces of fake news with high expected virality. This paper proposes a systematic approach to identify the type of content that enhances the diffusion of fake info and forecast their social spreading to predict which fake stories will go viral based on their text. We show that the impact of content features on virality is different for true and fakes news, and so the propagation of true and fake news is. Furthermore, we use machine learning nonparametric models to classify fake news according to their propagation level.