Educational data mining has received considerable attention in an effort to improve the quality of learning. Many data mining techniques have been proposed to extract the hidden knowledge from educational data. The extracted knowledge helps institutions to improve their respective methods and learning processes of their teaching and learning activities. All these improvements lead to an increase in the performance of students and with efforts to improve the quality of better learning output. In this study, the authors propose a new student performance prediction model based on data mining techniques with new data attribute behavior features, called student's behavioral features. These types of features are related to learner interactivity with the e-learning management system. The performance of the student predictive model is avoided by the classifier set, namely; Neural Networks, Naive Bayesian and Decision Trees. In addition, the authors apply the ensemble method to improve the performance of this classifier. The author uses Bagging, Boosting and Random Forest (RF), which is a common ensemble method used in the literature. The results obtained reveal that there is a strong relationship between learner behavior and academic achievement of the heirs. The accuracy of the proposed model using behavioral features is achieved up to 22.5% A 1% improvement compared to the results when removing such features and reaches up to 25% An 8% increase in accuracy using ensemble methods. By testing the model using newcomer students, more than 80% accuracy was achieved.

Performance Analysis, IBN Student Prediction Model, Data Mining, ensemble method