ASSESSING GRADUATION PREDICTIONS USING COMPARATIVE ANALYSIS OF CLASSIFICATION MODELS

Abstract

Data mining, also known as knowledge discovery in databases (KDD), involves extracting valuable, previously unknown information from large data volumes. This field is gaining significant importance in the educational sector, particularly within universities. This paper aims to predict students' final year grades using classification-based data mining techniques, assessing the performance of three algorithms – Naïve Bayes, J48, and SVM – to improve educational quality. By comparing these classification algorithms, we can evaluate their current efficiency and effectiveness. Various performance measures are utilized to compare the results from these classifiers. Our findings indicate that the J48 classifier achieves the highest accuracy among the tested classifiers, making it a valuable tool in predicting student graduation outcomes.