Sentiment analysis of students feedback: A study towards optimal tools

Educational Institutions attempts to gather feedback from students' to study their sentiments towards courses and instructors and to enhance the performance of the instructors. Basically, such feedbacks are gathered at the end of the semester with the use of survey forms. However, this technique is very tedious, slow and time consuming. With the advent of social media, especially Facebook, the collection of feedback become easier through Facebook pages and groups. But, analyzing those feedbacks is equally challenging. This paper addresses those problems and uncovers the best model for analyzing those feedbacks with the use of machine learning techniques such as Support Vector Machines (SVM), Maximum Entropy (ME), Naive Bayes (NB), and Complement Naive Bayes (CNB) and applying neutral class. And, found SVM as the highest performer with an accuracy of 97% by applying different preprocessing and feature extraction techniques and avoiding neutral class, which outperform state-of-art work by 2%.