Predictive analytics model for students' grade prediction using machine learning. A review

Abstract

Forecasting the academic achievement of students is a vital undertaking in the analysis of educational data, providing the opportunity to improve learning results and teaching methods. This paper investigates an array of machine learning (ML) methods for predicting student grades, offering a thorough examination of cutting-edge methodologies. The research centres on three primary projects that utilise supervised, unsupervised, and reinforcement learning techniques, namely Logistic Regression, Linear Discriminant Analysis, Random Forest, and Neural Networks. A comparative examination demonstrates that Neural Networks often surpass other models in terms of their predictive accuracy, adaptability, and capacity to represent intricate, nonlinear relationships in educational data. Notwithstanding issues like overfitting and the requirement for extensive training data, Neural Networks exhibit considerable promise for wider applications in personalized learning and academic prediction. The findings emphasize the profound influence of machine learning in educational environments, promoting the use of advanced algorithms to facilitate data-driven decision-making in academics.