Design and Analysis of Reinforcement-Learning and Graph- Based Curriculum Sequencing for Higher Education with OULAD and KDD Cup 2010 Datasets

DOI

10.2991/978-94-6239-674-6_25How to use a DOI?

Keywords

Reinforcement Learning; Graph Neural Networks; Personalized Learning Pathways; Curriculum Sequencing; OULAD; KDD Cup 2010; Deep Q-Network; Educational Data Mining; Intelligent Tutoring Systems; Adaptive Learning

Abstract

The research paper will come up with a hybrid system that combines the two models, Reinforcement Learning (RL) and Graph-Based Curriculum Sequencing in order to serve the needs of students in higher education institutions by personalizing the learning process. By using two open datasets, Open University Learning Analytics Dataset (OULAD) and KDD Cup 2010 (Khan Academy), the model will maximize the curriculum progression, and the prediction of engagement and adaptive content recommendation. The conceptualization views the educational process as a Markov Decision Process (MDP) whereby the state of knowledge of each student is dynamically changed by the learning interactions. A Deep Q-Network (DQN) agent chooses the following best course module or exercise according to the feedback of performance, time on task and level of cognitive mastery. In order to improve relational reasoning, a Graph Neural Network (GNN) is built on top of the curriculum graph, with learning activities forming the nodes and conceptual, performance-based, and prerequisite dependencies forming the edges. The RL agent is proposing the latent representations of the GNN to balance between exploration (adding new concepts) and exploitation (strengthening weak points). Comparative analyses with baseline sequence model and collaborative filtering recommender prove to be better with regard to knowledge retention ( +18), completion rate ( +12), and adaptive satisfaction index ( +15). The paper adds a scalable and interpretable contour of intelligent tutoring system, which is consistent with the objectives of 5.0 personalized education. The open-source datasets guarantee the reproducibility and extensions to multi domain applications are easily possible. The next step of development is to incorporate the emotional analytics and multimodal engagement capabilities to achieve an even more personalized approach towards learning in both hybrid and online learning environments.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Sachin Kumar Verma
AU  - Anil Kumar Gupta
AU  - Dilip Kumar
AU  - Gyanendra Veer Singh
AU  - Chandra Shekhar Verma
PY  - 2026
DA  - 2026/05/28
TI  - Design and Analysis of Reinforcement-Learning and Graph- Based Curriculum Sequencing for Higher Education with OULAD and KDD Cup 2010 Datasets
BT  - Proceedings of the International Conference on Sustainable Computing and Artificial Intelligence (ICSCAI 2025)
PB  - Atlantis Press
SP  - 286
EP  - 298
SN  - 2352-5401
UR  - https://doi.org/10.2991/978-94-6239-674-6_25
DO  - 10.2991/978-94-6239-674-6_25
ID  - Verma2026
ER  -