Cardiovascular Diseases (CVDs) are one of the leading causes of death. One major worry is Sudden Cardiac Arrest (SCA) caused by heart arrhythmia. Cardiac arrhythmias result from faulty electrical conduction or impulse production in the heart, affecting heart shape or heart rate. SCA is more likely in people with a history of stroke or cardiovascular risk. So regular monitoring of cardiac activity is necessary. Detecting arrhythmias is also important for adequate medication and preventing cardiac failure. An Electrocardiogram (ECG) gives critical information about heartbeats. An ECG is an electrical signal that represents the action potentials of various heart tissues. A Holter monitor is a portable ECG recorder that can record the heart's electrical activity for extended periods. Even for a skilled physician, analyzing a lengthy ECG record for aberrant rhythmic changes may be taxing. Thus, due to its efficacy and resilience, computer-aided diagnostics is crucial in arrhythmia detection. Arrhythmia detection follows class identification on an ECG. Heartbeat categorization is, therefore, a critical step in detecting arrhythmia. This thesis' major goal is to collect and analyze meaningful ECG information for successful heartbeat categorization. To do so, pick elements that disclose hidden information in ECG rhythms. Due to the nature of the ECG data, we used nonlinear decomposition methods to extract features. This thesis' primary contribution falls under three areas. The first category develops “class-oriented” heartbeat categorization. Using Ensemble Empirical Mode Decomposition (EEMD) and support vector machines, ectopic and bundle branch block heartbeats are distinguished from normal beats. It is tested on two databases in both noisy and quiet circumstances. The second category effectively addresses class imbalance using an enhanced EEMD, data level sampling methods, and an AdaBoost classifier. Our goal is to categorize “subject-oriented” heartbeats. To broaden the applicability of the suggested strategy, the same heartbeat groups are categorized in an “interpatient” system. This is vital in real-time applications.
Keywords: Classification, Class Imbalance, Nonlinear Decomposition, Arrhythmia, ECG signal.
Silan Software is one of the India's leading provider of offline & online training for Java, Python, AI (Machine Learning, Deep Learning), Data Science, Software Development & many more emerging Technologies.
We provide Academic Training || Industrial Training || Corporate Training || Internship || Java || Python || AI using Python || Data Science etc