<![CDATA[The evaluation of students’ academic performance is a fundamental aspect of any educational institution, playing a critical role in shaping students’ academic journeys and institutional decision‑making. However, this process presents signi icant challenges, particularly when dealing with large student populations. Traditional methods of result evaluation often lead to inef iciencies, delays in processing, and increased workload for institutions. With the rapid advancements in information technology and arti icial intelligence, automated systems have revolutionized student performance assessment,making the process faster,more accurate, and less labor‑intensive. Machine learning has emerged as a powerful tool in this domain, enabling the prediction of student performance through techniques such as regression and classi ication. While these models provide valuable insights, their effectiveness largely depends on accuracy. Achieving high accuracy in grade prediction remains a signi icant challenge, as even slight inaccuracies can lead to misclassi ication, affecting students’ academic outcomes. To overcome these limitations, ensemble learning methods have proven to be highly effective. These techniques combine multiple models to enhance predictive performance and reduce errors. This study focuses on evaluating various ensemble methods, including random forest, bagging, boosting, and extreme gradient boosting, to determine the most reliable approach for predicting student performance. A comparative analysis was conducted to assess the accuracy and ef iciency of these models using key evaluation metrics. The results indicate that extreme gradient boosting out performed other models, achieving the highest accuracy in predicting student grades. This research highlights the importance of ensemble learning in academic performance assessment andunderscoresits potential to improve decision‑making in educational institutions.]]>
Copyrights © 2025
