The Design and Evaluation of an Affective Tutoring System

Abstract

Emotion recognition techniques in an affective tutoring system enhance the affective interface of an intelligent tutoring system. This study attempts to improve interest in learning by recognizing the emotional states of students during their learning processes and providing adequate feedback. Both the emotional recognition system and the tutoring strategy module are developed. The digital arts learning content module is then designed, along with the emotional feedback mechanism via the HCI design of interactive avatars. Next, the entire system is integrated and assessed via a two-stage evaluation. Analysis results indicate that the proposed affective interaction design can enhance learner motivation and interest, thus improving their learning performance. System satisfaction and student performance are evaluated first by prototype evaluation, in which experts combine a system usability scale (SUS) questionnaire and heuristic evaluation. Triangulation evaluation is then performed, including qualitative and quantitative approaches involving observations, questionnaires, and interviews. The following questions are raised using the above methods: Is the usability of ATS effective?; Are ATS users satisfied with the learning system?; Is the interactivity of ATS attractive to users?; Does ATS increase the learning motivation of digital art students?; and Does ATS vary in its impact on self-conscious learning achievement for users from different domains? Evaluation results demonstrate a high satisfaction among users with respect to system usability. Furthermore, in addition to its attractiveness to users, ATS increases learning motivation and self-conscious learning achievement.