Scholarlink

<p>Electrocardiograms (ECG) are a valuable tool for the prediction of heart disease (HD), arrhythmias, and cardiovascular disorders, but most of the methods proposed in the literature have several limitations, including denoising, heartbeat segmentation, feature extraction, and classification. Diagnosing cardiac illness is challenging, and it is crucial to extract successfuldetailsthrough the large number of patient ECG recordings. To improve the HD prediction using advanced ECG signal analysis, this research designs a Meta-Learning and Deep Convolutional Recurrent Neural Network with Foraging Phero Trap Optimization (ML-DCRNN-FPTO) model, which combines meta-learning approach with Deep Learning (DL) architectures that are specially made to quickly adapt to new assignments with limited information. The Foraging Phero Trap Optimization (FPTO) algorithm is used to optimize the learning process, improving BioSignality (BS) feature extraction and model performance to predict the accurate region on HDThe model demonstrated remarkable accuracy of 95%, precision of 96%, F1-score of 95%, and recall of 95% with the MIT-BIH Arrhythmia Dataset. Similarly, the proposed model shows consistent improvement on the PTB-XL ECG Dataset, achieving 95% recall, 95% accuracy, 96% precision, and 95% F1-score, respectively.</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

Meta-Learning and Deep Convolutional Recurrent Neural Network with Foraging Phero Trap Optimization for Heart Disease Prediction

  • 期刊论文

  • 发表日期：

    2025年8月19日

查看全文
  • Mohammed Irshad,
  • Syed Mohamed E.

摘要

Electrocardiograms (ECG) are a valuable tool for the prediction of heart disease (HD), arrhythmias, and cardiovascular disorders, but most of the methods proposed in the literature have several limitations, including denoising, heartbeat segmentation, feature extraction, and classification. Diagnosing cardiac illness is challenging, and it is crucial to extract successfuldetailsthrough the large number of patient ECG recordings. To improve the HD prediction using advanced ECG signal analysis, this research designs a Meta-Learning and Deep Convolutional Recurrent Neural Network with Foraging Phero Trap Optimization (ML-DCRNN-FPTO) model, which combines meta-learning approach with Deep Learning (DL) architectures that are specially made to quickly adapt to new assignments with limited information. The Foraging Phero Trap Optimization (FPTO) algorithm is used to optimize the learning process, improving BioSignality (BS) feature extraction and model performance to predict the accurate region on HDThe model demonstrated remarkable accuracy of 95%, precision of 96%, F1-score of 95%, and recall of 95% with the MIT-BIH Arrhythmia Dataset. Similarly, the proposed model shows consistent improvement on the PTB-XL ECG Dataset, achieving 95% recall, 95% accuracy, 96% precision, and 95% F1-score, respectively.