Using Classification Data Mining Techniques for Students Performance Prediction

The big problem of retired or drop out students is about academic achievement. The educational institution needs to follow up the advising system since advisers should guide the planning curriculum to their advisees. The data mining techniques are applied to anaylze the performance of the students and to impart the quality of education in the educational institutions. This paper focuses on classification models for applying in Education Data Mining. The classification models are applied to identify the suitable subject to the science students. The experiment is set to improve the student performance which comparing the performance of five classification models and then predicting the appropriated academic achievement in each major. To examine the experiment, we used 17,875 academic achievements within 483 students. Four measures; precision, recall, f-measure, and accuracy are evaluated the models. For the result, the best accuracy is Gradient Boosted Trees: GBT at 92.41% and F-measure value equals to 84.59%.