NAÏVE BAYES ALGORITHM BASED ON FORWARD SELECTION ON THE GRADUATION PREDICT OF TIME

The number of students who do not pass on time, becomes a separate paradigm for institutions of higher education that must be handled properly, because in addition to affect the performance and lecturer ratio in the institution also brings a bad impact for the college. Therefore, timely graduation calculation for students is an important element in strategic decision making for an institution of higher education. Limitations in this research is on the comparison of calculation accuracy using Naive Bayes algorithm with calculation using Naive Bayes algorithm based on Forward Selection. The calculation result using Naive Bayes algorithm has 97,69% precision and accuracy on calculation using Naive Bayes-based Forward Selection algorithm produce better accuracy value that is 97,92%, so it can be concluded that the use of Naive Bayes Algorithm based on Forward Selection has the highest accuracy value with a precision 98,92%, and becomes the best model in problem solving in timely graduation predictions for students.