An automated course feedback system using opinion mining

This paper presents our experimental work on designing an automated web-based course feedback system that uses opinion mining to compute the overall feedback score. The system collects user responses through an online form. The online feedback form consists of a set of binary and graded response questions as well as a free-form input text box where a student can write about his overall experiences with the course. Every student is encouraged to write in detail about his experience. The text entered by the student in free form input text box is then used for opinion mining to label the feedback as `positive' or `negative'. We have used an unsupervised SO-PMI-IR based approach for opinion mining. The final feedback of a course is thus a combination of an overall score (computed on various binary and graded questions) and a `positive' or `negative' label to mark the student's overall evaluation. We evaluated the system for more than 20 courses with a total of about 1000 reviews.