Scholarlink

<p>Accurate and timely diagnosis of heart disease remains a critical challenge in modern healthcare. Traditional machine learning approaches, such as support vector machines (SVM) and Random Forests, have shown some success in predicting heart disease risk. However, these methods often struggle to capture complex, nonlinear relationships within the data, potentially leading to suboptimal diagnostic accuracy. To address these limitations, this study proposes a novel approach utilizing a deep learning architecture based on Transformer networks. The Transformer’s attention mechanism allows the model to dynamically weigh the importance of different patient features, capturing intricate interactions between various risk factors. Experimental results demonstrate the superior performance of the proposed Transformer-based model compared to traditional machine learning algorithms. The model achieved significantly higher accuracy (99%), F1-score, and AUC metrics on a comprehensive dataset of patient records. This improvement suggests that the Transformer architecture can effectively leverage the inherent complexity of the data to enhance the accuracy and reliability of heart disease diagnosis.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Toward precision cardiology: a transformer-based system for adaptive prediction of heart disease

  • 期刊论文

  • 发表日期：

    2025年5月2日

查看全文
  • Fatma M. Talaat,
  • Wesam F. Aly

摘要

Accurate and timely diagnosis of heart disease remains a critical challenge in modern healthcare. Traditional machine learning approaches, such as support vector machines (SVM) and Random Forests, have shown some success in predicting heart disease risk. However, these methods often struggle to capture complex, nonlinear relationships within the data, potentially leading to suboptimal diagnostic accuracy. To address these limitations, this study proposes a novel approach utilizing a deep learning architecture based on Transformer networks. The Transformer’s attention mechanism allows the model to dynamically weigh the importance of different patient features, capturing intricate interactions between various risk factors. Experimental results demonstrate the superior performance of the proposed Transformer-based model compared to traditional machine learning algorithms. The model achieved significantly higher accuracy (99%), F1-score, and AUC metrics on a comprehensive dataset of patient records. This improvement suggests that the Transformer architecture can effectively leverage the inherent complexity of the data to enhance the accuracy and reliability of heart disease diagnosis.