Predicting Student Academic Performance using Support Vector Machine and Random Forest

The use of machine learning and data mining in the educational field to predict student performance, known as educational data mining, which has always been an important study area. An early prediction of student performance may help the responsible entities to provide solutions to the students with low performance. Student performance in the final exam could be affected by many factors (e.g., previous assignment grades, social life, parents’ job, and absence frequency). This paper aims to predict student academics performance to enhance the performance of educational organizations to get better academic results of their students. In this paper, classification algorithms and techniques applied were Support Vector Machines (SVM) and Random Forest (RF). Binary classification and regression techniques have been applied with both SVM and RF. In this work, we predict the final grade of mathematics course and Portuguese language course, the dataset consists of 369 and 649 records, respectively. The experimental results for both SVM and RF algorithms applied to both datasets showed that the accuracy in the case of binary classification achieves a superior accurate prediction reach to 93%, while in regression, the lowest RMSE is 1.13 in case of RF.