Unsupervised ECG Analysis: A Review

Abstract

Electrocardiography is the gold standard technique for detecting abnormal heart conditions. Automatic detection of electrocardiogram (ECG) abnormalities helps clinicians analyze the large amount of data produced daily by cardiac monitors. As the number of abnormal ECG samples with cardiologist-supplied labels required to train supervised machine learning models is limited, there is a growing need for unsupervised learning methods for ECG analysis. Unsupervised learning aims to partition ECG samples into distinct abnormality classes without cardiologistsupplied labels-a process referred to as ECG clustering. In addition to abnormality detection, ECG clustering has recently discovered inter and intra-individual patterns that reveal valuable information about the whole body and mind, such as emotions, mental disorders, and metabolic levels. ECG clustering can also resolve specific challenges facing supervised learning systems, such as the imbalanced data problem, and can enhance biometric systems. While several reviews exist on supervised ECG systems, a comprehensive review of unsupervised ECG analysis techniques is still lacking. This study reviews ECG clustering techniques developed mainly in the last decade. The focus will be on recent machine learning and deep learning algorithms and their practical applications. We critically review and compare these techniques, discuss their applications and limitations, and provide future research directions. This review provides further insights into ECG clustering and presents the necessary information required to adopt the appropriate algorithm for a specific application.

XplainScar: Explainable Artificial Intelligence to Identify and Localize Left Ventricular Scar in Hypertrophic Cardiomyopathy from 12-lead Electrocardiogram

“…Unsupervised ECG clustering has been successfully employed to overcome challenges facing the ECG supervised learning classification, for instance, to resolve the imbalanced data problem and low-level automation of patient-specific ECG classifiers. 42 Motivated by these applications, we propose a novel yet simple unsupervised approach to reduce inter-patient variability among HCM patients while increasing separation between the Scar and NoScar groups of patients. Our method partitions patients into groups consisting of several sub-clusters each containing patients with similar ECG patterns (low inter-patient variability), where the number of patients belonging to the Scar or NoScar class dominate the other class (high separation between classes).…”

Fuzz-ClustNet: Coupled fuzzy clustering and deep neural networks for Arrhythmia detection from ECG signals

“…Moreover, it also requires substantial expertise to interpret the ECG signals. This has motivated researchers over the years to propose more accurate and automated techniques to improve the effectiveness and efficiency of ECG signal analysis [7]. Most existing work in the enhancement of arrhythmia detection from electrocardiograms is classified as parametric feature-based and signalprocessing based.…”

Non-contact voltage measurement of residential cables based on internal parameter conversion and centering probe assistance

“…As can be seen from equation (2), to reconstruct cable voltage V l through sensor output voltage V o , specific values of C l , C e , C n and C s need to be specified. Where, C s is the structural capacitance of the sensor, which is a fixed value and can be obtained through the digital bridge.…”

“…This excellent characteristic has made non-contact voltage measurement technology a research hotspot in different fields in recent years. For example, acquisition of weak biological potential [1][2][3], non-invasive load monitoring [4,5], overvoltage measurement [6,7], motor status monitoring [8,9], partial discharge monitoring [10].…”