<![CDATA[Based on observations and existing data in the UHAMKA Informatics Engineering Study Program, the number of students who do not graduate on time (8 semesters) in each generation will cause an accumulation of the number of students, lack of classrooms, and lack of parking space. One of the ways to increase student graduation on time is to predict from begin which students have the potential to didn’t graduate on time, so that preventive action can be taken by the study program management or faculty. Prediction can be done using data mining by utilizing data from students who have graduated. The data mining method used in this study is Naive Bayes using gender attributes, achievement index from semester one to semester four and semester one to semester four credits. The Naive Bayes algorithm will be made several models and the highest accuracy value will be sought from the model. The model evaluation uses K-fold Cross Validation and the prediction results will be used by the academic supervisor to evaluate students whose prediction results are unsatisfactory. The model with the best results is the 3rd model with an accuracy rate of 80.19%, a recall of 80.26%, precision 92.75% and F-Measure 86.05% which will be used for implementation in the student graduation prediction application.]]>
Copyrights © 2020
