Using data mining techniques to extract students' attitudes toward e-learning

Abstract

The rapid expansion of e-learning platforms, where students can share their opinions and express their thoughts, has become a rich source of data for opinion mining and sentiment analysis. This study aims to develop an effective model for predicting students' attitudes about e-learning, with a focus on mining opinions that indicate positive or negative sentiments. The study was implemented in two stages. The first stage aimed to discover the most popular platform used in e-learning at the University of Mosul to collect the largest amount of data through comments posted within the platforms, also to identify trends in students' opinions towards e-learning. The results show that the focus of both lecturers and students revolved around well-known platforms such as Google Classroom and Google Meet, both of which had relative importance (45.33% and 42.29%, respectively). The second stage uses a machine-learning algorithm on the data collected to determine the impact of e-learning on students. Also, two feature selection approaches, information gain (IG) and CHI statistics, were explored and enhanced in addition to hidden Model Markov (HMM) and support vector machine (SVM)-based hybrid learning strategy. As a result, an opinion mining method was used to assist developers in improving and promoting the quality of relevant services.