Application of the C5.0 Algorithm for Student Graduation Integrated with Academic Information Systems

Abstract

A series of processes that must be passed by each student, including completing a predetermined number of courses, carrying out practical fieldwork, proposal seminars, and final project seminars, all of which are carried out towards graduation. However, not all students can graduate on time according to the study period taken, so sometimes it causes a pile of students who do not graduate according to their graduation period. These students’ graduation time cannot always be detected early, leading to delays in graduates. To overcome this, a technique of the C5.0 algorithm forms a classification tree is proposed to predict student graduation. The C5.0 algorithm has a pruning technique that can trim the decision tree to be simpler, shorter, and boosting is used to measure the level of percentage accuracy. From the test results by applying the C5.0 algorithm using 300 students data from the class of 2019 (with 210 for training and 90 for data test), the percentage of recall, precision, accuracy, and error values were about 94.28%, 80.48%, 88.89%, and 11.11%, respectively.