International Research Journal of Education and Technology

Educational institutions generate vast amounts of data related to student demographics, academic records, attendance, and online learning activities. Analyzing this data using machine learning techniques can help predict student performance and identify students at risk of academic failure. Educational Data Mining (EDM) has gained significant attention as it enables institutions to extract meaningful insights from educational datasets. This research presents a comparative analysis of several machine learning algorithms for predicting student academic performance. Algorithms including Decision Tree, Random Forest, Support Vector Machine (SVM), Logistic Regression, K-Nearest Neighbors (KNN), and Gradient Boosting are evaluated using educational datasets.

The study focuses on identifying the most accurate algorithm for predicting student outcomes based on academic and behavioral attributes. The models are evaluated using several performance metrics including accuracy, precision, recall, F1-score, and ROC-AUC. Previous studies have demonstrated that ensemble learning algorithms often outperform traditional algorithms due to their ability to handle complex data relationships. The results of this study can assist educational institutions in identifying at-risk students and implementing timely interventions to improve academic performance.

Key Words:  Educational Data Mining (EDM), Student Performance Prediction, Machine Learning Algorithms, Decision Tree, Random Forest, Support Vector Machine (SVM), Logistic Regression, K-Nearest Neighbors (KNN), Accuracy, Precision, Recall, F1-Score, ROC-AUC.