Implementasi Algoritma Naive Bayes Classifier Menggunakan Feature Forward Selection Dalam Memprediksi Ketepatan Masa Studi Mahasiswa

Abstract

Students are one of the important pillars in the life cycle of a higher education institution. One indicator of the success of a study program can be seen from the accuracy of the student's study period. The accuracy of a student's study period refers to the time schedule that the student must achieve from entering the study program to graduating, according to the time span determined by the university. At the Diploma Three (D-III) level, it is said to graduate on time if you have completed the equivalent of three academic years of study and not graduate on time if you have completed more than three academic years. Study. This research aims to influence the quality of study programs, so that it is used as an assessment criterion for accreditation by (BAN-PT). The results of this research are presented in accordance with the research conducted. The data used in this research is academic data for students of the Informatics Management Study Program (D-III) Class 2020 – 2022. The categories used are gender, major, IPS 1 to IPS 4, credits, and GPA. The test results using 129 training data and 40 test data obtained an accuracy of 97.50% using the Naive Bayes method with forward selection. There were 12 students who made late predictions and 28 students made correct predictions. So it can be stated that the Naive Bayes process model is suitable for use in determining good decision results in terms of prediction and classification.