Enhancing Student Performance Prediction in Higher Education: A Data-Driven Approach

Student performance is a crucial factor in higher education institutions, as admission to a high-quality institution often relies on a good academic record. The primary objective of this paper is to predict student performance by considering their personal and academic achievements. By identifying poorly performing students, teachers can offer timely guidance and support to improve their academic outcomes. However, predicting student performance becomes challenging due to the processing of a large amount of data, including both numerical and non-numerical values. This work aims to determine the most effective prediction algorithm and identify the key variables in student data to enhance student performance and success rates through classification techniques using educational databases from both universities and schools. The proposed system leverages this data to predict a student's next year's result (GPA). The dataset used for this experiment is locally obtained from third-year students, encompassing their core subject results (6 subjects, 2 laboratories), and personal details. The methods employed in this work include the Multiple Linear Regression, Naive Bayes and Decision Tree. The extracted factors impacting the results will help students prepare better in advance. The highest accuracy achieved in this prediction is 88.44%, bringing significant benefits to students, teachers, and educational institutions.