Recomendação de estratégias pedagógicas baseada em emoções e perfis de personalidade em ambientes virtuais 3D

Abstract

Emotions are key part of the process of teaching and learning in general, directly or indirectly. Students with an unfavorable emotion to learning tend to have difficulties in absorbing the content that is suggested at that time. This work proposes an module that is able to detect and classify those emotions and, through a second logic module based on facts, suggests educational recommendations using Personality ProĄles in order to develop the student affective proĄle in Virtual Learning Environments (VLE). To computationally validate this work, Ąrst, it was observed which technique would present better results to detect the Action Units and then, which algorithm would show better results for this classiĄcation. In this last case, the decision tree showed itself as an effective method with an average rate of 86,4%. This method was evaluated by images and videos to ensure its applicability. After this process, it was used the Personality ProĄles along with the Case-Based Reasoning to suggest pedagogical recommendations through emotions association and the personality proĄles during the process of teaching and student learning. In this way, the student is being monitored all the time and, when a harmful emotion is detected, these instructions proposed in this work are executed in order to provide educational recommendation and offer him encouragement and/or motivation according to his individual characteristics. This brings the student closer to his emotions and improves learning signiĄcantly.