Implementation of Data Mining To Predict Period of Students Study Using Naive Bayes Algorithm

Abstract

The quality of universities, especially study programs in Indonesia is measured based on accreditation conducted by BAN PT. According to BAN PT the quality is measured based on 7 main standards, one of them is Student and Graduate. One of the problems that still be the subject of discussion related to student failure is about the students who graduated not on time. Students graduating not on time are students who can not complete their studies in accordance with the provisions of time given. The existence of a graduate student is not timely of course cause problems and potentially drop out that affect the quality of education and accreditation. A system that predicts students' graduation is required by evaluating their learning outcomes. The timeliness of graduating students can be done with data mining techniques to find graduation patterns of students who have graduated which then used as a basis to predict students' graduation in the next year. This study showed that Naïve Bayes was able to classify the correct data testing on average by 86.16% and 13.84% error. In addition, other information obtained from the data testing used that the students who entered from the PMDK Pass graduated on time as much as 40%, other paths graduated on time by 26.7%, and pass filter exam on time 13.3%.