Performance Comparison of Data Mining Classification Algorithms on Student Academic Achievement Prediction

Abstract

Academic achievement is one of the benchmarks of student success in carrying out the learning process. Grade Point Average (GPA) is a reference for universities in determining student academic achievement. For universities, academic achievement can be an indicator of determining the success of the learning system and can improve the image of the university. This study aims to determine the prediction of academic achievement results of Pamulang University students with Naive Bayes, C4.5 and KNN, and to determine the comparison results of Naive Bayes, C4.5 and KNN algorithms in predicting the academic achievement of Pamulang University students. The algorithms compared in this study are Naive Bayes, C4.5 and K-Nearest Neighbor (KNN) algorithms, using the factors of gender, age, faculty, regional origin, work status, organisation participation, type of school origin, distance of residence, and parents' profession as artibut. The results of this study show that the KNN algorithm is the algorithm with the greatest accuracy rate of 56.25%, followed by the Naive Bayes algorithm and the C4.5 algorithm with an accuracy rate of 50.00%.