The Development of Models to Forecast Numbers of New StudentsUsing the Classification Rules with Decision Tree Technique

Abstract

Advertisement and admission of target applicants, who will be new undergraduate students,will help to reduce costs, promote or recruit off- campus as well. The decision tree Technique can be used to develop models to forecast numbers of new students in the classification rules for accuracy. It helps to decide on the form and place in admission of new students to meet the target even more. The seven models developed by using the rules of classification techniques, decision trees were built and tested with different sampling methods: 3 models of k-fold cross-validation, 3 models of percentage split, and a model of training set and test set. The study result for forecasting new students via model developed by training set and test set method is higher performance than model developed by other methods. It was shown that the efficiency was 94%, the precision was 94.3%, the recall was 94% and the F-measure was 93.7%. Thus, this model is accurate in forecasting numbers of new students using the classification rules with decision tree technique.