One of the leading causes of death worldwide is heart disease. Therefore, in order to slow the rising death rate, early diagnosis of heart disorders is essential. The electrocardiogram (ECG) is a commonly used diagnostic tool for a variety of cardiac conditions, including irregular heartbeats (arrhythmias). It is quite challenging to identify the abnormal ECG signals' properties, nevertheless, due to their non-linearity and complexity. Furthermore, manually verifying these ECG signals could take a lot of time. To get over these restrictions, researchers have developed a quick and precise classifier that performs better than other well-known classifiers at simulating a cardiologist's diagnosis in order to distinguish between normal and pathological ECG signals from a single lead ECG signal. Analyzing and processing ECG data is a crucial step in the diagnosis of cardiovascular diseases. The main goal of this chapter is to understand the classification of healthy and sick people through popular machine learning-based methods.
TopIntroduction
Heart problems are currently the world's biggest cause of death (Zipes, 2000). The ECG is essential for studying cardiac physiology). The Electrocardiogram (ECG) records the electrical activity of the heart; the signal's amplitude is small, ranging from 0.4 to 3 mv. The ECG's low magnitude value highlights how important it is to choose the best filtering approach before processing any further data in order to do biological research. Surface electrodes are commonly used to record ECG signals. Three lead configurations are offered: bipolar, unipolar, and improved unipolar. The sources of noise include muscle noise, power line interference, baseline drift, also referred to as low frequency noise, and electromagnetic interference from other equipment are the sources of noise in ECG readings (Appathurai et al., 2019).
An irregular heartbeat, or cardiac arrhythmia, is the result of the heart beating too rapidly or too slowly. Misaligned electrical impulses that synchronize heartbeats are the cause of it. Sudden cardiac death can be caused by certain major arrhythmia conditions. Consequently, the main objective of an ECG study is to accurately detect potentially lethal arrhythmias so that the appropriate treatment can be given to maintain life. Over the past few decades, many methods for automatic ECG beat classification have been described. In this study, we present a thorough analysis of the state-of-the-art methods for identifying cardiac arrhythmia from ECG signals. It consists of the machine learning, feature extraction, and signal decomposition approaches utilized in autonomous detection and decision-making (Sahoo et al., 2020).
The heart muscle pumps blood to the vital organs, which is essential to human survival. Congestive heart failure (CHF) is the inability of the heart to adequately pump blood throughout the body without raising intracardiac pressure. Arrhythmia, confusion due to reduced cerebral blood flow, and lung and peripheral congestion, which results in breathing problems and swollen extremities, are among the symptoms. Coronary artery disease, myocardial infarction, and medical co-morbidities such diabetes, hypertension, and renal disease can all cause heart damage and reduced myocardial function. The prevalence of CHF is increasing worldwide. It is one of the leading causes of death and affects millions of people globally. Consequently, it's critical to appropriately diagnose, track, and manage (Jahmunah et al., 2019).
As per the World Health Organization, cardiovascular diseases are responsible for thirty percent of deaths worldwide and pose a significant burden on society (Fuster & Kelly, 2010). Therefore, early patient identification of individuals who are at risk and a fuller understanding of the illness causes are essential for enhancing diagnosis and therapy. Doctors frequently utilize electrocardiograms (ECGs) in hospitals to capture the electrical signal traveling from the skin to the heart. Because of this, the ECG can detect various structural or electrical abnormalities in the heart, which can help with the diagnosis of cardiac diseases. Manually analyzing large amounts of ECG data can be time-consuming and difficult. As a result, powerful computational methods are needed to maximize the information extracted from large ECG datasets (Obermeyer & Emanuel, 2016). The multiplicity of ECG forms and their clinical applications necessitate the use of a number of computer techniques to address this demand.
The electrocardiogram (ECG) is the most often utilized diagnostic method for evaluating the electrical activity and anatomy of the heart. The spectrum of clinical conclusions that can be made from ECG data could be greatly expanded by concurrent advancements in processing power, machine learning techniques, and the availability of vast amounts of data while retaining interpretability for use in medical decision-making (Tison et al., 2019).
The electrocardiogram, often known as an ECG or EKG, is a non-invasive electrical trace of the heart that is obtained from the skin. German is where the term “ECG” originates. It is known as electro-kardiographie in German. The Dutch physician Einthovan created ECG in 1902, and over the course of roughly 10 years, his enormous contributions to clinical research led to the full realization of the technique's clinical potential (Krikler, 1987).