Manuscript received September 28, 2024; revised November 1, 2024; accepted November 22, 2024; published February 14, 2025
Abstract—Accurately predicting student performance in e-learning environments is a significant challenge that is essential for personalizing education and enhancing learning outcomes. This study examines the effectiveness of machine learning techniques in forecasting learner success within e-learning ecosystems, using the Open University Learning Analytics Dataset (OULAD). We conducted a comparative analysis of four machine learning algorithms—Random Forest, Logistic Regression, Support Vector Machine (SVM), and Linear Discriminant Analysis (LDA)—employing comprehensive data pre-processing and feature engineering methods. The Random Forest algorithm outperformed the others, achieving a 91% accuracy rate in classifying student outcomes into “Distinction,” “Pass,” and “Fail” categories. Despite this high accuracy, differentiating between certain performance classes, especially “Distinction” and “Fail,” remained challenging, highlighting the complexity of student performance metrics in online learning contexts. These results demonstrate the potential of machine learning, particularly the Random Forest algorithm, as a valuable tool for enhancing predictive analytics in e-learning systems. The study contributes to the optimization of educational technologies by indicating that refined predictive models can lead to more effective, data-driven interventions and personalized learning pathways.
Keywords—machine learning, predictive analysis, e-learning, student performance, learning analytics, random forest algorithm
Cite: Fatima Ezzahraa EL Habti, Mustafa Hiri, Mohamed Chrayah, Abdelhamid Bouzidi, and Noura Aknin, "Enhancing Student Performance Prediction in e-Learning Ecosystems Using Machine Learning Techniques," International Journal of Information and Education Technology, vol. 15, no. 2, pp. 301-311, 2025.