Data mining using filtering approaches and ensemble methods

Abstract

To develop a prediction paradigm, ensemble methods such as boosting based on the heuristic system can be used. Compared to using individual classifiers, the prediction results using ensemble learning techniques are usually more accurate. In this study, several ensemble techniques were discussed to obtain comprehensive knowledge about key methods. Among the various ensemble methods, a boosting mechanism has been implemented to predict student achievement. Because the ensemble method is considered an actual event in a prediction and classification, the boosting technique is used to develop an accurate predictive pedagogical model. The utilization of this boosting technique is based on the nature of each method proposed in educational data mining. By using the ensemble method and filtering approach, the predictions of student performance showed a substantial increase.