Support Vector Regression for GPA Prediction

Peneliti

Kania Evita Dewi, S.Pd., M.Si Nelly Indriani Widiastuti, M.T.

Deskripsi/Abstrak

This study aims to predict student GPA. This research began by collecting data. The features used in predicting GPA are semester 1 and semester 1 IP grades. The process of GPA prediction uses SVM regression, Linear Regression and Simple Linear Regression. Based on testing with normalized data, the smallest error is obtained by the SVM regression method with Kernel RBF which is equal to 0.1505. Whereas by using standardized data, the smallest error is obtained by using the SVM regression improve method with the Kernel RBF, which is 0.1487. Based on this research, in order to obtain prediction results that are closer to the actual values, it is better to standardize the data first and to predict the process using the SVM Regression Improve method using the Kernel RBF