An Explainable Educational Data Mining System for Predicting Student Academic Performance

Students' educational accomplishments are vital as they are a key indicator of their overall academic journey. This study proposes a Machine Learning (ML) enabled approach to categorize students' academic performance with the explainability of the proposed EDM system. We utilized standard ML models like Decision Tree, Random Forest, and Boosting models such as Gradient Boosting and Extreme Gradient Boosting (XGB) for multi-class classification, where XGB outperforms other models with 83% accuracy, precision, and recall. The proposed EDM approach demonstrates superior performance compared to existing works, offering a more effective approach for anticipating and enhancing student academic outcomes. The explainability results demonstrate that changes in curriculum, courses, and the status of paying tuition fees can bring a positive result in the student's academic performance. Future research focuses on designing a recommendation system to help the students justify their current status and get personalized support to enhance academic performance in real time.