An Intelligent Academic Performance Prediction Model Using Hybrid Machine Learning Techniques

Predicting student academic performance accurately is vital to inform academic interventions and institutional decisions. However, existing educational prediction approaches may rely on individual machine learning (ML) models with limited predictive capabilities. In view of these limitations, this study introduces a hybrid model for Student Academic Outcome Prediction (SAOP) based on Random Forest and Extreme Gradient Boosting (RF-XGBoost). The proposed framework combines the merits of both algorithms by aggregating their probabilities with weights for better classification results. Experimental results demonstrate that the hybrid model achieves 96.5% accuracy (ACC), 96.3% precision (PRE), 96.2% recall (REC), and a 96.7% F1-score (F1). Compared with Support Vector Machine (SVM), Convolutional Neural Network (CNN), and Stacking Ensemble models, the proposed approach achieves the highest reported ACC, outperforming them at 87.08%, 90.49%, and 91%, respectively. The results suggest the proposed framework can be a viable solution to predict students' academic outcomes and inform data-driven educational planning.