<![CDATA[This study aims to develop a robust machine learning framework for predicting student learning mastery in Informatics subjects. The research employs a supervised learning approach using assessment-based features derived from student academic records. Due to limitations commonly found in real educational data, including imbalance and data leakage risks, synthetic data generation and feature engineering were applied to support controlled experimentation. Several classification models were implemented to evaluate the stability and consistency of the proposed framework. The results indicate that the models were able to consistently distinguish between students who achieved learning mastery and those who did not. The comparable performance across different modeling approaches suggests that the predictive capability is driven by the methodological design rather than dependence on a specific algorithm. This study demonstrates that machine learning can provide a reliable and interpretable tool to support data-driven evaluation and early intervention in Informatics education.]]>
Copyrights © 2025
