ECG Dimensionality Reduction using PCA and Feature Abstraction Expending ICA Built with Power Spectral Estimation

Electrocardiography, which is more often referred to as an electrocardiogram, is a way of monitoring the electrical activity of the heart over a period of time using electrodes that are put on the skin. These electrodes are able to pick up on the minute electrical changes in the skin that occur as a direct consequence of the electrophysiological pattern of depolarization and repolarization that occurs in the heart muscle with each beating. It is typically done to detect heart problems. The primary focus of this study is on extracting features from ECG wave data using Independent Component Analysis and Principal Component Analysis, both of which were developed with the help of a non-parametric power spectral estimation method. In a real-time setting, doctors establish diagnoses using criteria based on a minimal set of features extracted from an electrocardiogram. Parameters like signal amplitudes, distances, slopes, and morphology fall within this category. The primary goal of feature extraction is the real-time extraction of ECG signal components, which provides the necessary data for specialists or doctors to do additional analysis. Feature Extraction is largely accountable for reducing the effort and time needed to effectively describe large data sets. When analyzing complicated data, the amount of variables that must be taken into consideration in order to complete the analysis is often the biggest bottleneck. Analyses performed on large datasets place additional demands on the system's memory and processing power. Dimensionality reduction primarily entails defining a system with fewer, more manageable variables through the use of vector-based reduction techniques. Features are extracted from the reduced variables that can be used in the processing. Therefore, PCA is used for dimensionality reduction and ICA constructed with non-parametric power spectral estimation is used for feature extraction to accomplish classification in this work.