Multi-Surrogate-Enhanced Deep Neural Network for Dynamic Medical Decision Support
NTU Journal of Pure Sciences · 12.07.2026 · 5 (2) · 85-96
DOI: 10.56286/8n21kv40 ↗Kurz gefasst
Fachartikel in der Zeitschrift NTU Journal of Pure Sciences, online erschienen am 12.07.2026. Verfasst von Warqaa Hashim Alkhaled. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Medical Decision Support Systems (MDSSs) have become indispensable tools in modern healthcare, assisting physicians and healthcare professionals in disease diagnosis, treatment planning, patient monitoring, and clinical decision-making. The rapid growth of artificial intelligence (AI) and deep learning has significantly enhanced the capabilities of these systems by enabling accurate analysis of complex medical data. However, conventional deep learning models often require extensive computational resources, large training datasets, and long processing times, limiting their application in real-time clinical environments. To address these challenges, this study introduces a Multi-Surrogate-Enhanced Deep Neural Network (MSEDNN) model that integrates surrogate modeling techniques, including Radial Basis Functions (RBF) and Support Vector Regression (SVR), with deep neural network architectures. The proposed framework improves feature extraction, reduces computational complexity, enhances real-time adaptability, and supports personalized clinical decision-making across multiple medical tasks. The performance of the proposed model was evaluated using several benchmark medical datasets involving diagnostic classification and medical imaging applications. Experimental results demonstrated that the MSEDNN model achieved an average classification accuracy of 96.5%, with a 97.0% sensitivity and an F1-score of 95.6%, indicating excellent diagnostic performance and reliability. Furthermore, the framework attained 97.2% accuracy on medical imaging datasets while maintaining a low inference latency of 140 ms, making it suitable for time-sensitive clinical applications. The incorporation of surrogate-based preprocessing also reduced model training time by approximately 25%, significantly improving computational efficiency without compromising predictive performance. These findings demonstrate that the proposed MSEDNN framework provides an effective, accurate, and computationally efficient solution for intelligent medical decision support. By combining surrogate modeling with deep learning, the model offers enhanced diagnostic accuracy, faster response times, and improved adaptability, highlighting its potential for integration into next-generation healthcare systems and precision medicine applications.
Zitieren
APA
Alkhaled, W. H. (2026). Multi-Surrogate-Enhanced Deep Neural Network for Dynamic Medical Decision Support. NTU Journal of Pure Sciences, 5(2), 85-96. https://doi.org/10.56286/8n21kv40
Vancouver
Alkhaled WH. Multi-Surrogate-Enhanced Deep Neural Network for Dynamic Medical Decision Support. NTU Journal of Pure Sciences. 2026;5(2):85-96. doi:10.56286/8n21kv40
BibTeX
@article{Alkhaled2026,
author = {Alkhaled, Warqaa Hashim},
title = {Multi-Surrogate-Enhanced Deep Neural Network for Dynamic Medical Decision Support},
journal = {NTU Journal of Pure Sciences},
year = {2026},
volume = {5},
number = {2},
pages = {85--96},
doi = {10.56286/8n21kv40}
}
EUVIMED