<![CDATA[Education at this time is an important requirement in facing the demands of an increasingly advanced era in technolo-gy. To compensate this, the existing educational standards in universities must also be improved, this is a bit much affect the pattern of teaching from universities that produce qualified graduates who can compete in the world of work later and indirectly give a positive impact on the university itself. Qualified graduates are of course not only depending on the role of a university but also majors and quality of education as long as students are still in high school / vocational school also plays an important role. Results of the on-time graduation rate prediction research can be used as an information to im-prove the quality and optimization of the education system but it requires a maximum degree of accuracy. This research predicts on time graduation rates by conducting analysis using data mining classification techniques. Naïve Bayes algo-rithm that are used for this research will be discussed as a reference in conducting research. The author performs a series of different experimental scenarios / cross validation to perform comparisons that can give a difference in the level of ac-curacy gained from this research. The results of this research indicate that with the addition of Cross Validation testing scenario there is an increase of 2% accuracy of the test.]]>
