Emotion Detection and Opinion Mining from Student Comments for Teaching Innovation Assessment

Abstract

Students can provide their opinions, comments, or suggestions about a course, course instructor, study environment, and available resources using the course evaluation at the end of every semester. This helps the course professors and other college authorities make appropriate changes or continue a particular approach to get the best experience in classrooms. These course evaluations are in both quantitative and qualitative forms. In quantitative feedback the evaluation is performed in terms of measurable outcomes and include a Likert-type scale to capture the level of agreement and disagreement. In qualitative feedback the students can convey their feelings, opinions or suggestions about the course, the course instructor, or their overall thoughts/comments towards the course. The qualitative feedbacks provide freedom for the students to express their honest thoughts on a course. The data collected in the qualitative form provides deeper insight into a student’s emotional state. In this work we focus on mining the qualitative student feedbacks and analyzing the student sentiments. We also analyze the efficiency of Light Weight teams and Flipped Classroom approach which are Active Learning methods. Results show that the implementation of these Active Learning methods is linked with increased positivity in student emotions.