Skip to main content

Abstract

Nowadays, disinformation is everywhere. On many occasions, the use of disinformation can make some citizens change their opinion regarding a particular issue. This can lead an influenceable citizen to start to tend towards extremist thoughts and become radicalized and even become an extreme activist or cyber activist. To mitigate this, most research focuses on detecting false or extremist information in order to remove it from the media as soon as possible. What if it is too late and that false information has already penetrated a citizen's mind? In this case, it is necessary to analyze the behavior of citizens who consume this information; and the simplest and most effective way to do this is by analyzing their behavior regarding their community on social networks. The objective is to develop a system that using natural language processing, graph analysis, and machine learning techniques can detect anomalous behaviors related to Online Social Networks and Media information. Thus, the study of Social network graphs will lead to detecting different communities or groups of users and the peculiar users within each community. The system must be able to dynamically gather users’ behavior to capture its possible peculiarity fluctuations over time. Forecasting techniques will be needed so that the system will be able to anticipate future behavior to create appropriate mitigation strategies.