Citation

Ghani, Abdul and Javad, Hooman Oroojeni Mohamad and Mushtaq, Shougfta (2026) A Hybrid Stacked Ensemble of Deep Learning and Machine Learning Models for High-Accuracy Diabetes Prediction. Information Systems Engineering and Management, 86. pp. 771-785. ISSN 3004-958X

Text A Hybrid Stacked Ensemble of Deep Learning and Machine Learning Models for High-Accuracy Diabetes Prediction _ Springer Nature Link.pdf - Published Version Restricted to Repository staff only Download (813kB)

Abstract

Diabetes mellitus is a commonly encountered chronic disease that causes substantial healthcare problems all over the world. Its increasing frequency emphasizes the persistent demand for active and effective timely detection and management strategies. To address this global issue, this research presents a hybrid stacked ensemble model that combines deep learning, machine learning and optimized feature engineering techniques by using the Pima Indian Diabetes Dataset (PIDD). The proposed model achieved the accuracy 96.75%±2.19, precision 95.96%±3.77, recall 94.80%±4.12, F1-score 95.32%±3.13 and AUC 99.18%±0.65. These results substantially outperform the outcomes of existing deep learning approaches, which stated precision of 95.22%, accuracy of 98.07%, Recall of 95.52% and F1-score of 99.29%. The enhanced performance of the proposed system highpoints its robustness, reliability and the potential for integration into the clinical decision-support systems, it eventually contributing to amended patient care and practical diabetes management.

Item Type:	Article
Uncontrolled Keywords:	Deep learning and machine learning
Subjects:	Q Science > Q Science (General) > Q300-390 Cybernetics
Divisions:	Faculty of Computing and Informatics (FCI)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	01 Oct 2026 02:06
Last Modified:	01 Oct 2026 02:07
URII:	http://shdl.mmu.edu.my/id/eprint/16749

Downloads

Edit (login required)