Unleashing Deep Learning for Academic Success: A Survey of Predictive Models in Educational Data Mining

V. Loganayaki, N. Balakumar

Abstract: Educational Data Mining (EDM) is a fast-emerging area of influence in Educational Informatics which aims to enhance educational outcomes by applying foundations and techniques of data mining to educational data. The prediction of students' academic performance is one of the many applications of EDM, which is of great significance in providing timely academic support to at-risk students once detected in their early stages. This review examines the application of Deep Learning (DL) techniques for student performance prediction within EDM. The increasing availability of educational data and the limitations of traditional statistical and machine learning approaches have encouraged the adoption of deep learning models capable of automatically learning complex feature representations. The study reviews major deep learning architectures, including Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Long Short-Term Memory networks (LSTMs), attention-based models, graph neural networks (GNNs), and hybrid frameworks. A comparative analysis of datasets, evaluation metrics, strengths, limitations, and predictive performance is presented. The findings indicate that deep learning models generally achieve high predictive accuracy, particularly when temporal and behavioural learning patterns are incorporated. The review also highlights challenges related to interpretability, computational complexity, and data dependency. Finally, future research directions are identified to support the development of reliable, scalable, and practical student performance prediction systems.

Keywords: Educational Data Mining, Student Performance Prediction, Deep Learning, Learning Analytics, Convolutional Neural Networks, Long Short-Term Memory Networks, Educational Artificial Intelligence

How to Cite?: V. Loganayaki, N. Balakumar, "Unleashing Deep Learning for Academic Success: A Survey of Predictive Models in Educational Data Mining", Volume 15 Issue 6, June 2026, International Journal of Science and Research (IJSR), Pages: 674-680, https://www.ijsr.net/getabstract.php?paperid=SR26612152545, DOI: https://dx.dx.doi.org/10.21275/SR26612152545

Download Citation: APA | MLA | BibTeX | EndNote | RefMan