Student Performance Prediction Using AI-ML and Business Intelligence

Portugal still trails behind in Europe despite improvements in recent decades in the educational level of its population. This is mostly because of Portugal's high rates of student failure, especially in basic subjects like Portuguese and Mathematics. Similar difficulties exist in India, where there is a chronically high percentage of student failure, particularly in science and math. This study uses business intelligence/data mining methodologies to address these problems by using real-world data, such as demographics, grades, and school-related aspects. In order to test four DM models and three input options, including prior grades, both nations use binary/five-level classification and regression tasks to model student performance. When past school period grades are taken into account, the results demonstrate increased prediction accuracy. The academic performance of students is subject to other factors outside their prior scores, including parental employment and education, alcohol consumption, and absenteeism. The findings of this study indicate the possibility of developing more effective techniques for predicting student outcomes, with the goal of improving the quality of education and optimizing the management of school resources.