Combination of Case Based Reasoning with Nearest Neighbor and Decision Tree for Early Warning System of Student Achievement

Student achievement is one of the main focuses to increase university credibility. An early warning system is needed to prevent more risks. The early warning systems (EWS) of student achievement has been possible with a combination of case based reasoning (CBR), k-nearest neighbor (K-NN), and decision tree. CBR is used to obtain a solution that stores knowledge so that it can predict student achievement. This research is combining the Case Based Reasoning, K-Nearest Neighbor (K-NN), and Decision Tree (DT) methods for the prediction of student achievement that applied in the early warning system. The attributes of an early warning system of student achievement are genders, distances of residence, ages, high schools, majors, and grade point average (GPA) for six semesters. The results show that accuracy rate is 60.5% of 55 data in the early warning system of student achievement and a model CBR for Early Warning System of Student Achievement.