An affective sensitive tutoring system for improving student’s engagement in CS

Abstract

With the growing popularity of online teaching and tutoring, there are many attempts to enhance students’ learning experience during the lecture. This paper presents an animated tutoring system for improving student engagement using nonverbal cues, including students’ facial expressions. The system can (1) capture students’ facial expressions in the scenario; (2) identify various facial expressions, including anger, disgust, fear, sadness, happy, and surprise; and (3) provide feedback to students based on students’ facial expressions. To evaluate the tutoring system, we predicate the student engagement using support vector machine with the captured information, and measure students’ engagement using students’ academic performance, i.e., in-system exercise, quizzes, and exams. Our empirical study shows that the student performance using the level 2 animation is 10% and 20% high then levels 1 and 0, respectively.