Prediction of Student Performance by Using Machine Learning Techniques

Each person’s learning process is distinct and diversified. Therefore, using the right teaching techniques is crucial to improving student performance and academic achievement. Numerous educational institutions, including schools and universities, make an effort to anticipate their students’ academic performance in order to improve the effectiveness of the educational systems. In educational technology and the recent learning systems, discovering useful hidden patterns in learner data is significantly valuable. In this paper, several machine learning classifiers were applied such as Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Support Vector Machine (SVM), and Naive Bayes (NB), to predict the student’s performance. Furthermore, a hybrid model was used to reduce data dimensionality and improve the overall performance of the classifiers used. A real dataset obtained from the UCI machine learning repository is used to evaluate the proposed hybrid model. The reported results show the performance of CNN-RNN outperforms other methods with an accuracy equal to 89.1%.