Citation

Abuowaida, Suhaila and Alshdaifat, Nawaf and Mashagba, Hamza A. and Aziz, Azlan Abdul and Alzoubi, Alaa and Roslee, Mardeni and Alias, Mohamad Yusoff and Mahmud, Azwan (2026) An Enhanced Explainable Hybrid Framework for Modeling Metaverse-Based E-Learning Acceptance Among Students with Disabilities: Integrating SEM, Neural Networks, and Interpretable Machine Learning. Metaverse, 7 (2). ISSN 2810-9791

Text 8449-58814-1-PB.pdf - Published Version Restricted to Repository staff only Download (1MB)

Abstract

While metaverse technologies offer numerous opportunities for inclusive learning for children with disabilities, there is limited empirical research on how these same children accept or reject such technologies; as such, there are few if any, established methodological frameworks, which can be used to explain the complex acceptance behavior of children with disabilities toward metaverse technologies. This research provides an advanced hybrid analytical framework that will extend beyond structural equation modeling (SEM) and artificial neural networks (ANN) by employing additional explanatory AI techniques, ensemble models, and clustering methods to analyze student acceptance data from children with disabilities. This study will be conducted using a quantitative cross-sectional survey design with a combination of five complementary analytical approaches): 1.) Structural equation modeling (SEM), 2.) Deep neural networks, 3.) Random forest and XGBoost ensemble methods to determine the most important features related to student acceptance, 4.) SHAP and LIME methods to understand why the various AI models developed were making predictions regarding student acceptance, and 5.) Unsupervised clustering to identify subgroups of students based on type of disability. Using a large dataset of over 847 students with disabilities from multiple school-based settings, this research demonstrated improved predictive power of the enhanced framework when compared to the use of single-method approaches, while maintaining sufficient levels of theoretical interpretability. Overall, the findings of this research indicate that the two primary acceptance factors for children with disabilities toward metaverse technologies are: 1.) Perceived optimism toward the benefits of the metaverse technology, and 2.) Social presence. Furthermore, the relative importance of each factor varied significantly depending upon the specific type of disability (cognitive vs. physical). This finding highlights the need for design teams to develop acceptance-enhancing strategies tailored to the needs of students with different types of disabilities. Additionally, the findings of this research demonstrate the potential of the explainable AI components of the framework to provide educators and developers with specific actionable insight into the factors driving acceptance decisions made by individual children. Therefore, this research contributes to the field of special education research an empirically grounded, methodologically sound, theoretically interpretable, and practically applicable framework for studying the acceptance of technology in inclusive educational environments, as well as the practical applications of developing acceptance enhancing features within metaverse platforms.

Item Type:	Article
Uncontrolled Keywords:	Metaverse, Explainable AI, Machine learning, SHAP, LIME
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines
Depositing User:	Ms Suzilawati Abu Samah
Date Deposited:	31 Jul 2026 07:08
Last Modified:	31 Jul 2026 07:08
URII:	http://shdl.mmu.edu.my/id/eprint/16428

Downloads

Edit (login required)