A Real-time ECG CTG based Ensemble Feature Extraction and Unsupervised Learning based Classification Framework for Multi-class Abnormality Prediction

Cardiovascular diseases (CVDs) are a leading cause of death worldwide. Early detection and diagnosis these can greatly reduce complications improve outcomes for high-risk individuals. One method detecting CVDs is through the use electrocardiogram (ECG) monitoring systems, which various technologies such as Internet Things (IoT), mobile applications, wireless sensor networks (WSN), wearable devices to acquire analyze ECG data early diagnosis. However, despite prevalence systems in literature, there need further optimization improvement their classification accuracy. In an effort address this challenge, novel heterogeneous unsupervised learning model real-time was proposed. The main goal work error rate accuracy system. This study presents framework multi-class abnormalities electrocardiograms (ECGs) using ensemble feature extraction technique learning. utilizes electrocardiogram-cardiotocography (ECG-CTG) system extract features from signal, then employs techniques enhance discrimination extracted features. used learning-based algorithm classify signals into different classes abnormalities. proposed evaluated on dataset results show that it effectively with high low computational complexity.