Abstract
[Author declined to make full text available]Cardiovascular diseases (CVDs) remain the leading cause of mortality globally, with developing and remote-area countries disproportionately affected due to limited access to healthcare infrastructure and a shortage of specialised medical professionals, such as senior doctors and cardiologists. The introduction of emerging technologies has enabled the collection of vast amounts of data, particularly in the healthcare sector. Harnessing this data, this research introduces a novel approach that utilises Machine Learning (ML) and Deep Learning (DL) techniques to develop an intelligent, scalable, and cost-effective diagnostic system for heart disease diagnosis.
By carefully identifying and incorporating critical diagnostic features, this study evaluates logistic regression (LR), decision tree (DT), random forest (RF), and neural network (NN) models, alongside a uniquely designed Convolutional Neural Network (CNN) architecture optimised for Electrocardiogram (ECG) signal analysis. Unlike many existing studies that limit feature selection to 14 attributes, this research incorporates additional key factors identified through in-depth medical analysis, improving the diagnostic system's precision and robustness. Stratified k-fold cross-validation is employed to ensure reliability and minimise bias, and the results are benchmarked against state-of-the-art methods.
The proposed system achieves high diagnostic accuracy, even when tested on small datasets, making it highly applicable in settings where extensive medical data or expertise is unavailable. This system has the potential to transform healthcare delivery in resource-limited settings by enabling frontline healthcare workers to diagnose CVDs accurately without the need for constant oversight by specialists. Furthermore, the technology is adaptable for integration into wearable devices and telemedicine platforms, extending its reach to underserved and remote populations.
This work highlights the importance of combining data-driven innovation with real-world healthcare needs, addressing differences in access to quality care and empowering developing nations to combat the growing burden of heart disease effectively. Ultimately, given the complexity of this research area, the study also explores open challenges and potential directions for future work.
| Date of Award | 3 Apr 2025 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Alexander Gegov (Supervisor), Matthew Poole (Supervisor) & Mohamed Bader-El-Den (Supervisor) |