Using Sentiment Analysis to Explore Student Feedback: A Lexical Approach

Abstract

Given the increasing abundance of online courses over the last couple of years, new forms of student feedback, which are less frequently used by teachers, have been generated in massive amounts. Nonetheless, extracting and processing this student generated content manually is costly and time consuming. In this respect, our objective in this paper is to propose a lexical-based approach that can predict the underlying sentiments of each student review, thus, enabling teachers to assess to what extent are students satisfied with the online learning resources and teaching practices. To enhance the performance of the proposed approach, a new education sentiment lexicon was built and incorporated into the model. After its implementation on a dataset that was extracted from the Web, this sentiment analysis lexical approach has proven to correctly predict the sentiment polarities of the great majority (i.e. 86.45%) of student feedback.