A design of machine learning-based adaptive signal processing strategy for ECG signal analysis

The human heart is categorized into four sections such as left and right atrium and left and right ventricle. The development of medical trained systems has increased the necessity for efficient new signal processing approaches to detect irregularities in order to diagnose heart-related disorders. The primary goal of this research is to provide medical treatments for people and hospital management systems. Observing and taking precautions for every human heart is an extremely fundamental part. The early prediction, is crucial for saving and giving the accurate attention to people about diet plans and way of life plans. Additionally, this is utilized to enhance the medical diagnosis and treatment of every affected patient. Here, to identify the heart-based problems Electrocardiogram (ECG) is utilized to analyze the electrical signal of the heart from the human body surface. Therefore, in this article, a novel Catboost-based echo State Network (CbESN) module is proposed to predict the human heart condition. Here, the standard datasets were utilized as implementation namely the PhysioNet 2017 database. Initially, the collected datasets are trained to the system then the preprocessing and feature extraction process take place. After that, classification is performed under different classifications such as normal and abnormal heartbeats to predict heart-based disease. Additionally, the developed CbESN has provided the finest outcomes of framework efficiency. Here, the proposed techniques achieve finest accuracy measure as 99%, precision measure is 99%, sensitivity is 99.1%, specificity is 98.02%, F-1 score is 97.23% and lower error rate as 0.016% for whole ECG signal processing. © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024.