Klasifikasi Prestasi Akademik Mahasiswa Berdasar Hasil Tes Potensi Akademik Menggunakan Support Vector Machine

Abstrak

This paper describes research that aims to classify student academic achievement based on the results of the Academic Potential Test (APT), using data containing APT scores and Grade Point Average (GPA) in semester 1 to semester 4. To get the best accuracy, some experiments were carried out by varying the fold, kernel, and multiclass. The folds used in the experiment are 5, 7, and 9 folds. The kernel used is a linear kernel, Gaussian Radial Basic Function (RBF), and polynomial. Meanwhile, the multiclass used in this experiment is one against one and one against all. The highest accuracy of 80% was obtained in the GPA classification in semester 4 using multiclass one against one and one against all, RBF kernel, and 9-fold cross validation