STUDENT GRADUATION PREDICTION SYSTEM BASED ON ACADEMIC AND NON-ACADEMIC (EQ) DATA USING C4.5 ALGORITHM

Abstract

The graduation profile is an important element for higher education accreditation standards. It reflects the performance of the adopted education system within a certain period. The better the profile graduation, the better the value of the accreditation. Some students are unable to complete their studies on time or even fail to complete their studies because they exceed the specified time limit, which is seven years, and it negatively affects institutions' accreditation. To prevent this from happening, it is necessary to know what obstacles that cause these students could not complete their studies on time. by knowing this information, prevention can be done for students who are potentially unable to complete their studies on time. The purpose of this study was to make a system that can predict the graduation timeline and the factors that influence it. The data used was graduation data from undergraduate students majoring in psychology from 2015 to 2017 at a university in Cimahi. The data had a total record of 461 students, 44 subject value attributes, 13 psychotest attributes, and class attributes. We generated the result by using decission tree method with C4.5 algorithm, which produces 90.32% accuracy. The depth of the tree can also influence the accuracy of the algorithm. This study also found that academic and non-academic (EQ) scores can affect students’ graduation time.