Using data mining to extract knowledge from student evaluation comments in undergraduate courses

Abstract

In this paper, we conduct a preliminary study about a newly established course evaluation survey given to the students by our institution. This survey contains several free-text questions which generates more qualitative but also voluminous unstructured feedback compared to the previous Likert scale question-based survey. Our aim is to apply data mining techniques to extract knowledge from these surveys, with the goal to help educators and administrators gain insight into student sentiment and views. Specifically, we apply text mining with classification to student comments regarding their perception of the teaching as a whole, in order to categorize them as positive or negative comments. Additionally, we apply text mining with association rule mining to various student comments to extract important key terms and associations of terms in these comments. Our preliminary results are encouraging and demonstrate the usefulness of employing data mining techniques to extract knowledge from the open-ended comments in our student surveys.