Improving e-learning with sentiment analysis of users' opinions

Abstract

E-learning has witnessed a great interest from the part of corporations, educational institutions and individuals alike. As an education pattern, e-learning systems have become more and more popular. It commonly refers to teaching efforts propagated through the use of computers in a bid to impart knowledge in a non traditional classroom environment. As a prerequisite for an effective development of e-learning systems, it is important to have certain knowledge about users' opinions and build an evaluation regarding them. Hence, an opinion mining method has been applied in this paper for the sake of helping the developers to improve and promote the quality of relevant services. Actually, three feature selection methods MI (Mutual Information), IG (Information Gain), and CHI statistics (CHI) have been investigated and advanced along with our proper HMM and SVM-based hybrid learning method. In fact, the experimental results have indicated that opinion mining becomes more difficult and challenging when performed for e-learning blogs. Moreover, we attempt to demonstrate that IG performs the best potential for sentimental terms selection and exhibits the best performance for sentiment classification.