A Machine Learning Based Approach for Student Performance Evaluation in Educational Data Mining

Astonishing progress in machine learning and data mining techniques has been achieved during the last two decades. Education should benefit from these improvements to discover about how people learn in different educational settings. The aim of the research is comparing artificial neural network (ANN) to random forest (RF) machine learning models for predicting performance of students based on their demographic and assessment information. After analyzing the Open University Learning Analytics Dataset (OULAD), we applied feature-engineering techniques, and then the two models were evaluated. Our results showed that the ANN model outperformed the RF model accuracy 91.08& to 81.35 &. ANN performs well on educational data and can be efficiently used for predicting student performance and in early warning systems.