همایش بین المللی هوش مصنوعی

صفحه اصلی / اولین همایش بین المللی هوش مصنوعی

Hybrid Deep Learning Models for Cardiovascular Disease Prediction: A Comprehensive Review of Convolution-Transformer Architectures

نویسندگان :

Ali Azimi Lamir¹ Masoud Bekravi² Babak Nouri Moghaddam³

1- Department of Computer Engineering, Ardabil Branch, Islamic Azad University, Ardabil, Iran 2- Department of Computer Engineering, Ardabil Branch, Islamic Azad University Ardabil, Iran 3- Department of Computer Engineering, Ardabil Branch, Islamic Azad University, Ardabil, Iran

کلمات کلیدی :

heart disease classification،hybrid models،convolutional neural networks،transformers،attention mechanisms

چکیده :

cardiovascular disease remains a leading cause of death worldwide, demanding innovative approaches to improve early detection and diagnosis. This study provides a comprehensive review of hybrid deep learning models that integrate Convolutional Neural Networks (CNNs) and Transformers for enhanced heart disease classification. It delves into the mathematical underpinnings of CNNs and Transformers, elucidating their distinct mechanisms for processing and extracting features from ECG signals. The paper examines various hybrid architectures, emphasizing their effectiveness in combining local and global feature extraction to achieve superior classification accuracy. Additionally, it discusses key challenges, including computational complexity, interpretability, and data constraints, while proposing potential solutions and outlining future research directions. Comparative analysis on benchmark datasets underscores the superior performance of hybrid models in cardiovascular disease prediction.