Analysis of Scholarship Consideration Using J48 Decision Tree Algorithm for Data Mining

Consideration of scholarships is a common occurrence in educational institutions such as in a university. The scholarship selection committees play an essential role in judgment, which must pay attention to considering issues efficiently. However, they may make mistakes because an applicant’s information is complicated. This research proposes a scholarship analytic for the award of a student scholarship at university by using Data Mining techniques. The study was designed with seven variables on 468 samples, which were only selected with complete attributes from 2,549 student documents by a decision tree, J48 and J48graft algorithm with percentage split method at 20%, 30%, and 60%, k-fold cross validation both 5-folds and 10-folds. The development model’s results found that the model created by a decision tree with the J48 algorithm and percentage split method at 66% is most effective, with the precision value at 77.35%. Therefore, we choose to model with the J48 algorithm by percentage split method at 66% to develop the web application, which is useful for students to assess themselves before applying and will decrease the committee’s workload for the assessment of student’s scholarship applications.