Identification of Fake Messages in The Media Based on Their Emotional Coloring

Abstract

The purpose of this research is to determine the feasibility of identifying fakes based on their emotional tone. The research objective is to develop a methodical approach that allows for the automatic identification of fakes without complex verification procedures using software tools. The method we propose does not involve revolutionary mathematical models but rather compiles existing developments in a specific sequence of procedures, which constitutes its scientific novelty. The advantage of this approach lies in its simplicity and relatively easy application, even for users with basic computer skills. To illustrate and facilitate replication, we provide a research flowchart and a link to the "Orange" program with open-source code. Specifically, the methodology involves calculating emotional probabilities according to Plutchik and classifying messages based on a Random Forest (RF) model, implemented within the mentioned program through widgets.