Predicting Student Performance Through Data Mining: A Case Study in Sultan Ageng Tirtayasa University

• Published in: Journal of advanced computational intelligence and intelligent informatics Vol. 27; no. 6; pp. 1159 - 1167
• Main Authors: Alfanz, Rocky, Hendrianto, Raphael Kusumo, Siagian, Al Hafiz Akbar Maulana
• Format: Journal Article
• Language: English
• Published: Tokyo Fuji Technology Press Co. Ltd 20.11.2023
• Subjects: Accuracy; Algorithms; Case studies; Classification; Colleges & universities; Control systems; Data mining; Decision trees; Educational objectives; Graduation rate; Informatics; Intelligence; Learning; Performance prediction; Students; Support vector machines
• ISSN: 1343-0130; 1883-8014
• DOI: 10.20965/jaciii.2023.p1159

Failure in compulsory subjects such as chemistry, calculus, physics, and basic control systems could hamper the graduation process of students. Thus, students must be successful in such obligatory courses. To address this issue, this study aims to predict student performance based on their learning outcomes using data mining techniques. In particular, we utilize decision tree (DT), k -nearest neighbor (KNN), support vector machine (SVM), and naive Bayes (NB) algorithms to predict student performance. The data for this study were gathered from the learning outcomes of students in the basic control systems course and subsequently modeled using binary and nine-level classifications. The experimental results showed that DT could perform better than KNN, SVM, and NB in the binary and nine-level classifications. Interestingly, the results of DT (i.e., the prediction values) are almost similar to those of the original values of the basic control systems course.