Development of a Smartphone Application and Chrome Extension to Detect Fake News in English and European Portuguese

Abstract

In a digital society, the truth portrayed by information is crucial in promoting education, security, and evolution. However, fake news raises a significant concern in that regard. Although there has been a continuous effort in the fight against fake news, it is still a multifaceted challenge in constant change as the menace renovates itself. Thus, in our approach, several machine learning and deep learning models were developed to obtain models that can detect fake content that appears online. The models can then be interfaced with users devices, namely in the form of browser extensions and smartphone applications. The classification models run on a cloud server and are accessible via web services. These models can detect fake news in English and European Portuguese, with a stronger focus on the latter, given the reduced number of projects in this specific field and language. Besides developing the first public dataset for fake news detection in European Portuguese through web scraping, the models achieved better performance than previous work while being trained with a significantly higher amount of data from a wider variety of sources.