The Prediction on the Students’ Graduation Timeliness Using Naive Bayes Classification and K-Nearest Neighbor

Abstract

Listen

Translate article

The college quality can be seen from the level of punctuality of student graduation. The Prediction on students’ graduation timelines can be used as one of the supporting decisions to evaluate students’ performance. Currently, the Medical Laboratory Technology study program of STIKES Guna Bangsa Yogyakarta does not have tools to predict the level of students’ graduation punctuality early yet. The purpose of this study is to evaluate the application of the Naive Bayes Classification and K-Nearest Neighbor algorithms with predictive modeling of student graduation period. This study applied the academic data from students of the Medical Laboratory Technology study program for the Academic Year (TA) 2015/2016 to 2018/2019. This study utilized an experimental approach by comparing the methods of the Naive Bayes Classification (NBC) and K-Nearest Neighbor (KNN) algorithms. The validation model uses 5-fold Cross Validation, while the evaluation model uses a Confusion Matrix. The results illustrated that the prediction with NBC in this case obtained an accuracy of 96.11%, precision of 82.11% and Recall of 100.00%. Meanwhile, predictions using KNN obtained accuracy of 97.68%, precision of 100.00% and Recall of 86.11%. Thus, KNN is an algorithm with an enhanced level of accuracy to solve the case of predicting the timeliness of students’ graduation of the Medical Laboratory Technology Study Program STIKES Guna Bangsa Yogyakarta