Automatic Student Performance Prediction System Using Data Mining Techniques

Educational institutions like universities and schools are now using Machine Learning to predict student performance. With the help of Educational Mining, analysts can now review and inform students about any risky patterns, and hence provide additional counselling or training to students. This work aims to develop an automatic student performance prediction system using the Student Performance dataset that predicts the grades of students. The dataset was pre-processed and analyzed to remove redundant attributes. Then, several classifiers like Support Vector Machine, Logistic Regression, Random Forest, Decision Tree, XGBoost, and K-Nearest Neighbour were experimented for the task of predicting the grade. Ensemble models using the combination of the individual classifiers stated earlier were also explored. The experimental results show that the ensemble model (Support Vector Machine + Decision Tree + K-Nearest Neighbour) achieved the highest accuracy of 90.57%. Additionally, the SVM classifier also exhibited the highest individual accuracy of 90%. The proposed method can be used by teachers and students in academic departments.