Mining Changes of Opinions Expressed by Students to Improve Course Evaluation

Abstract

Opinion mining can be used in many applications. In universities, students' opinions about courses can be considered as a significant informative resource to improve the effectiveness of education. Past works in this area focused on direct mining of students' opinions in regard to the courses. The aim of this paper is to develop a system which detects changes of students' opinions. Understanding such changes can help the management improve course evaluation in academic institutions. For course evaluation, knowing what is changing and how it has changed is crucial as they allow the management to provide the right course features such as teachers, contents, teaching materials and exams which need to satisfy the students' needs. In our work, we present a strategy for mining opinion changes based on the associative classification approach. Firstly, we collect opinions from students in two different semesters in regard to a specific course. Then, we extract rules using association rules. For this purpose, we detect and measure students' change of opinion from one semester to another. We describe types of opinions which can be detected by the students. Finally, we shed light on some of the examples which we have spotted from each type of opinion change.