KLASIFIKASI LAMA STUDI MAHASISWA FSM UNIVERSITAS DIPONEGORO MENGGUNAKAN REGRESI LOGISTIK BINER DAN SUPPORT VECTOR MACHINE (SVM)

Abstract

Graduation is the end result of the process of learning for studying in college. In attaining a normal S1 takes that for four years, but there are plenty of students who completed their studies beyond normal limits (over four years) and there is also less than four years. Older students study can be affected by many factors, among others, Graduation Achievement Index (GPA), gender, major, long pursued studies, scholarships, part-time, organizations, and university entrance pathway. In this study, the classification will be based on the status of a student's study time more than four years and less than or equal to four years. The method used to study the old classification of students with nominal data type is the method of Support Vector Machine (SVM) and will be compared with the Binary Logistic Regression method. Based on the research results of the binary logistic regression method, showed variables influencing the study period students are Subject and GPA with a classification accuracy of 70%. While the classification accuracy by using SVM highest classification accuracy using a linear kernel, polynomial and RBF reached 90%. Keywords: Older studies, Binary Logistic Regression, Support Vector Machine (SVM), Classification Accuracy.