KAJIAN PENERAPAN ALGORITMAC4.5, NEURAL NETWORK DAN NAÏVE BAYES UNTUK KLASIFIKASI MAHASISWA YANG BERMASALAH DALAM REGISTRASI

Abstract

Registration is a registration activities conducted in each semester. In an
educational institution, for the administrationis veryimportant. If an educational
institution having problems in the administration will be useless and can notlast long. Registration done old studentin each semester is the best way for educational institutions including the University Indraprasta PGRI, such as to be able to determine the number of classes that will be prepared at the beginning of the semestertuition. From the data known to thedecrease in the number of students who will be attending for the nextterm. Therefore,any factorthat causes many problems in the registration of students. Purpose of this researchis to create a classification problem or astudent whois not in the registration,in the study conducted comparison algorithm C4.5, naïvebayes and neural
network which is applied to the data in the registration troubled students. This study aimed to measure the accuracy of the comparativestudy of classification algorithms in 3 pieces that students have trouble registering. From the test resultsto measure the performance of the three algorithms using Cross Validation testing methods, Confusion Matrix and the ROC curve, it is known that Naive Bayes algorithm has he high estaccuracy value, ie91.57%, followed by C4.5 method with the accuracy of 91.43% and the lowest Neural Network is a method with a value of89.02% accuracy.

Keyword: Data Mining, Algoritma C4.5, Naïve Bayes, Neural Network.