Classifying Engineering Students Performance in Online Education with Machine Learning: Affective, Cognitive, and Behavioral Aspects

Abstract

In the rapidly evolving online learning environment, accurate prediction of student performance plays an important role in improving the quality of education and overall student outcomes. This study investigated the effectiveness of machine learning algorithms in classifying student performance in online courses based on affective, cognitive, and behavioural factors to develop more effective teaching strategies and interventions to support student success. A dataset of 485 engineering students who took Astronomy Physics at a private university in the fall and spring semesters of 2022-2023 was used to train and evaluate six machine learning algorithms: Support vector machines (SVM), K-nearest neighbours (KNN), Naive Bayes (NB) classifier, logistic regression, decision tree, and random forest. The random forest algorithm achieved the highest classification accuracy (87%), correctly classifying 87% of students into one of three performance categories: high, medium, or low. Moreover, the study determined that anxiety and expectations are the most influential factors in increasing student performance in online courses, while the least effective feature is social isolation. The findings suggest that Machine learning can efficiently categorize student performance in online courses, even with a limited set of features, enabling educators to enhance teaching strategies and interventions for better student support.