Affiliations:
1College of Computing and Informatics, Saudi Electronic University, Riyadh, Saudi Arabia
2Faculty of Engineering, Portsaid University, Port Said, Egypt
Aligning exam questions with course learning outcomes and linking them to program learning outcomes is often a time-consuming process that is prone to human error. This study examines the effectiveness of machine learning techniques for automatically mapping exam questions to Program Learning Outcomes (PLOs) and performance levels. A dataset of 414 multiple-choice questions was used to develop prediction models based on both joint and single-model architectures. The results show that the automated models achieved higher accuracy than human evaluations, indicating strong potential for the use of AI-based tools in educational quality assurance. The proposed approach can support academic institutions by automating assessment-related tasks, reducing faculty workload, and improving curriculum alignment. To the best of our knowledge, this study is the first to address the automated mapping of exam questions to program learning outcomes using machine learning methods.
Machine learning, Program learning outcomes, Exam question mapping, Educational assessment, Quality assurance
https://doi.org/10.21833/ijaas.2026.01.015
Alharthi, H., Alhargan, A., & Habib, M. (2026). Utilizing machine learning to align exam questions with program learning outcomes. International Journal of Advanced and Applied Sciences, 13(1), 144–153. https://doi.org/10.21833/ijaas.2026.01.015