Predicting high school graduates using Naive Bayes in State University Entrance Selections

Competition to entering the best and major state universities that are suitable for student interest is a problem for some high school graduates. There is three entrance selection system to the state university, they call National Selection for Entering State University (SNMPTN), Joint Selection for Entering State University (SBMPTN), and University Self Selection. SNMPTN becomes a favorite admission for students because it does not require written exams. SNMPTN is carried out before the other admissions are conducted, therefore the SNMPTN admission had very high competition. Students will be proud when they are accepted in the favorite state university with the major of their interest. This research was conducted to predict the high school students majoring in science to entering the state university via SNMPTN admission at first choice using Naive Bayes and SMOTE methods. SMOTE is used to overcome the imbalance class of data. The results showed that using SMOTE can increase the accuracy of the Naive Bayes method. The performance of Naive Bayes showed an accuracy rate of 90.1%, a precision of 52,23%, and the recall of 51,66%. While, the performance of Naive Bayes combined with SMOTE method gave 5.39% improvement of the accuracy rate, 43.3% of the precision, and 43.83% of recall.