Methodology for Automatic Identification of Emotions in Learning Environments

Abstract

. This paper presents a methodological proposal for automatic identification of emotions in educational environments using machine learning algorithms and physiological and behavioral signal acquisition technologies to identify relations between emotions and learning. Four of the main learning-centered emotions are considered [1]: engagement, boredom, confusion and frustration. It is proposed to make a fusion of data from two physiological and behavioral signal acquisition technologies with the objective of achieving the identification of emotions in the most precise manner. Therefore, considering the stages of the proposed methodology, the first of them is presented and the design of the experiment that will be executed for data collection. The development of an appropriate database with elements belonging to a learning environment for the study of emotions is an essential task.