ARCHIVES

Educational Data Mining (EDM) has emerged as an effective approach for extracting meaningful insights from educational datasets to support academic decision-making and improve student learning outcomes. Early prediction of student academic performance enables educational institutions to identify at-risk students and implement timely intervention strategies. This study proposes an explainable ensemble learning framework for predicting academic performance using a publicly available Kaggle student performance dataset. The dataset was preprocessed by handling missing values, removing duplicate records, encoding categorical variables, standardizing numerical features, and eliminating the ExamScore attribute to prevent target leakage. Four supervised machine learning algorithms, namely Logistic Regression, Decision Tree, Random Forest, and Support Vector Machine, were developed and evaluated using Accuracy, Precision, Recall, and F1-score metrics. Experimental results revealed that the Random Forest classifier achieved the highest predictive performance with an accuracy of 90.58%, followed by the Decision Tree model with 88.17% accuracy. Logistic Regression and Support Vector Machine produced comparatively lower performance, indicating that ensemble tree-based methods are more effective in capturing the complex relationships among academic, demographic, and behavioral features. The proposed framework provides an interpretable and reproducible approach for academic performance prediction and can assist educators in identifying students requiring additional academic support. The findings demonstrate that ensemble learning techniques can serve as effective decision-support tools in educational institutions and contribute to improved student success through data-driven interventions.