IMPLEMENTASI ALGORITMA K-NEAREST NEIGHBOUR UNTUK PREDIKSI KETEPATAN KELULUSAN

Abstract

In an education system, students are an important asset of a college and therefore, it is important to pay attention to the percentage of students who graduate on time. However, there is an imbalance between the inputs and outputs of the completed students. Students who enroll in large numbers, but students who graduate on time compared to those who are late according to regulations are fewer. In this study, the author aims to apply the K-NN method using cross validation to predict student graduation rates at STMIK IKMI. The results of this study are in the form of models and evaluations of student graduation predictions, whether they graduate on time or not on time. Based on the results of the design, implementation, testing using the RapidMiner program for predicting student graduation using the k-NN method with Cross Validation resulting in an accuracy of 70.28%, an error of 29.78%, and AUC of 0.739   Keywords: Graduation, Student, K-NN, Cross Validation