This paper describes how to use machine learning for improving teaching methods through collected sentiments from students. In fact, students sentiment analysis is a promising research area that is used to improve education by monitoring students performance and enabling students and lecturers to address teaching and learning issues in the most beneficial way. In our research, we aim to propose a machine-learning system for improving teaching methods through sentiment analysis, utilizing comments of students in reviews websites. The proposed system aims to automatically classify and analyze the students positive or negative feelings towards the current teaching process. Several techniques and procedures commonly used in natural language processing for the features processing task are used in designing and developing the proposed student sentiment analysis system. A total of 4000 comments of students were collected from RateMyProfessors.com website and used in the experiments of the current study. We have applied three supervised machine-learning techniques on these comments: Multinomial Naive Bayes (MNB), MaximumEntropy(MaxEnt), and Support Vector Machines (SVMs). The performance of the mentioned classifiers is evaluated using accuracy, precision, recall, and F1-score evaluation metrics.
Digital Object Identifier (DOI)
Omran, Thuraya; T. Sharef, Baraa; Hadjar, Karim; and Subramanian, Suresh
"Machine Learning for Improving Teaching Methods Through Sentiment Analysis,"
Applied Mathematics & Information Sciences: Vol. 14:
2, Article 15.
Available at: https://digitalcommons.aaru.edu.jo/amis/vol14/iss2/15