Tutorial Intervention’s Affective Model Based on Learner’s Error Identification in Intelligent Tutoring Systems

Intelligent Tutoring Systems (ITS) environments have the ability to adapt to each learner’s individual needs and thus provide immediate and personalized instructions, both in content and in form. This personalization can consider several aspects, such as the interaction, the level of knowledge, the error, and the affective state of the learner aiming to improve the teaching strategies. One of the strategies is the possibility of presenting tutorial interventions when verifying an error made in solving an exercise, or when detecting that the learner is unmotivated or frustrated. These tutorial interventions can improve teaching methods in order to improve performance and the level of knowledge acquired by the learner. In this sense, this research presents a model that allows the automatic presentation of tutoring interventions based on identification of mathematical error kind committed by the learner, in addition to inferring his affective state. Experiments were carried out in a real learning environment, using the proposed model implemented in a fraction game. In general, the results presented indicate that personalized tutorial interventions favor greater engagement and motivation of learners and improvement in learning outcomes.