Development of a Model for Predicting Students’ Achievement

In this study, data mining was implemented to find the common variables such as “gender”, “age”, “GPA first semester”, “GPA second semester”, “organization activity”, “part time job”, “living place”, “family income”, “father education”, “mother education” influencing the grade point of average (GPA) score of the 3nd-semester student at the Department of History Education. Three methods, including logistic regression (LR), decision tree (DT), and support vector machine (SVM) were employed. According to the validation results, the best algorithm method is found in the model developed by decision tree with the Accuracy 0.96 and all models provide sufficient data since the AUC value for all classes is greater than 0.5. This finding proved that above variables are linked to student achievement. As a result, concern to those aspects is critical for improving academic performance.