Scholarlink

<p>Sudden cardiac arrest (SCA) must be promptly diagnosed in order to avoid serious cardiac problems and death. Cardiovascular activity is measured by medical professionals using electrocardiograms (ECGs), which are vital bio-signals. Although ECG markers have been studied recently for the early diagnosis of SCA, the majority of existing techniques can only predict SCA up to 15&#xa0;min before onset, which is frequently insufficient to start medical treatment or hospitalization. In order to identify SCA at least an hour before it manifests, this study suggests an automated method that employs statistical analysis of ECG signals. Based on statistical parameters—such as skewness, kurtosis, and standard deviation (SD)—derived from the R-R intervals of the ECG, the approach uses many classifiers to differentiate between persons who are at risk of SCA and normal subjects. In tests using a dataset of 38 individuals from the Physionet repository, the suggested strategy outperformed other existing techniques, obtaining a maximum classification accuracy of 89.47% of the Naive Bayes classifier when predicting SCA an hour before its onset. This strategy shows how effective the recommended method is at detecting SCA early.</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

Early Sudden Cardiac Arrest (SCA) Prediction Using Different Classifiers and Statistical Features Analysis of ECG Signals

  • 期刊论文

  • 发表日期：

    2026年6月2日

查看全文
  • Prakash Banerjee,
  • Kousik Dasgupta,
  • Sourav Banerjee,
  • Abir Chattopadhyay

摘要

Sudden cardiac arrest (SCA) must be promptly diagnosed in order to avoid serious cardiac problems and death. Cardiovascular activity is measured by medical professionals using electrocardiograms (ECGs), which are vital bio-signals. Although ECG markers have been studied recently for the early diagnosis of SCA, the majority of existing techniques can only predict SCA up to 15 min before onset, which is frequently insufficient to start medical treatment or hospitalization. In order to identify SCA at least an hour before it manifests, this study suggests an automated method that employs statistical analysis of ECG signals. Based on statistical parameters—such as skewness, kurtosis, and standard deviation (SD)—derived from the R-R intervals of the ECG, the approach uses many classifiers to differentiate between persons who are at risk of SCA and normal subjects. In tests using a dataset of 38 individuals from the Physionet repository, the suggested strategy outperformed other existing techniques, obtaining a maximum classification accuracy of 89.47% of the Naive Bayes classifier when predicting SCA an hour before its onset. This strategy shows how effective the recommended method is at detecting SCA early.