The Future of Education
AI Driven Prediction of Student Success in Engineering Mathematics: Integrating Course Performance, Academic Progression, and Mathematical Misconceptions
Jinendrika Weliwita, Higher Colleges of Technology (United Arab Emirates)
Mohammed Almasalmeh, Higher Colleges of Technology (United Arab Emirates)
Abstract
Early identification of students at risk of academic difficulties is a major challenge in engineering education. Mathematics, as a foundational component of engineering programs, plays a critical role in determining student progression and overall academic success. This study presents an Artificial Intelligence driven framework for predicting student success in engineering mathematics by integrating course performance data, academic progression patterns, and mathematical misconception indicators.
Student achievement data collected over multiple academic years were analyzed using statistical and machine learning techniques. Correlation and regression analyses were first employed to investigate relationships between foundational and advanced mathematics courses. The analysis was then extended through the development of AI-based predictive models trained on student achievement records, progression histories, and misconception assessment results. Model performance was evaluated using standard classification and prediction metrics. The results revealed significant positive correlations between foundational and advanced mathematics courses, confirming the importance of strong mathematical foundations. The AI models effectively identified students at risk of poor academic outcomes with high predictive accuracy. Feature importance analysis depicted that prior mathematics achievement, progression history, and misconception scores were among the strongest predictors of future performance. Furthermore, while mathematical misconceptions exhibited a strong relationship with immediate course achievement, their influence on later mathematics courses was reduced, suggesting that many misconceptions are progressively corrected through higher education. Analysis of underperforming students also demonstrated that a substantial proportion improved in subsequent mathematics courses, highlighting the value of timely intervention and academic support.
The proposed framework demonstrates the potential of AI and learning analytics to support evidence based decision making, personalized interventions, and early warning systems. The findings contribute to the growing field of AI in Education and provide practical strategies for improving student success, progression, and retention in engineering programs.
Keywords: Artificial Intelligence in Education, Learning Analytics, Student Success Prediction, Mathematical Misconceptions, Predictive Modeling, Educational Data Mining.
