Implementasi Algoritme Decision Tree (C4.5) dengan Optimize Weights (PSO) untuk Memprediksi Kelulusan Mahasiswa Tepat Waktu

Abstract

The implementation of quality education is a requirement for higher education institutions to produce quality students. Government regulations regarding the limitation of the maximum length of lectures that students can take are intended so that the institutions involved in the educational process have careful planning in the learning process. Good planning is expected to be able to accreditation value of higher education institutions and deliver student studies so that they can graduate on time. In order to achieve these results, aspects that are so influential on the optimal value of accreditation results are needed, one of which is that students graduate on time and play an importan role in determining the results of accreditation. Decision tree itself is a very simple algorithm and easy understanding, decision tree algorithm alonei is enough to produc good and optimal accuracy values. Therefore, the optimization method particle swarm optimization (PSO) with its advantages can increase the level of accuracy by removing irrelevant features. The results of the study using a student dataset of class 2016 - 2017 explain that the optimization ofparticle swarm optimization (PSO) can produce 92.36% accuracy and increase the accuracy of 01.05%, with the decision tre method C4.5 with an accuracy rate of 91.31%. Furthermore, the T-test testing process is carried out on the decision tree algorithm C4.5 and the decison tree algorithm C4.5 optimization of particle swarm optimization (PSO) with the final result alpha = 0.635 where the results are less significant, it is said to be significant if the test results are below alpha = 0.050.