Comparison of Decision Tree, Naïve Bayes and KNearest Neighbors for Predicting Thesis Graduation

Achmad Solichin

Abstract


Thesis is one of the evaluations of learning for students. In Universitas Budi Luhur (UBL), especially in the Informatics Department, the thesis is one of the requirements for graduating students to obtain a Bachelor of Computer degree. In each semester, the number of Informatics Department students who take thesis is around 200-300 students. The problem that is still faced is that student graduation in the thesis is not optimal. Student failures in the thesis are allegedly related to several technical and nontechnical factors. In this study, an analysis using data mining algorithms was carried out to determine the factors that influence student graduation in the thesis. The dataset obtained from the Informatics Department students who took a thesis in the 2016/2017, and 2017/2018. In order to obtain the right classification method, this research was tested with three classification methods, namely Decision Tree, Naïve Bayes, and k-Nearest Neighbors (kNN). The results of the comparison of the values of accuracy, precision, and recall indicate that the kNN algorithm has advantages, so this method is chosen to predict graduation. In this study also developed an application for predicting graduation of students' thesis by applying the kNN classification method. The test results showed an accuracy of 78.20%, precision of 80.32%, and recall of 96.49%. This research is expected to be useful for improving the service quality of student thesis

Keywords


data mining; student thesis; kNN; Naïve Bayes; decision tree; comparison;

Full Text: PDF

Refbacks

  • There are currently no refbacks.