Analysis of Automatic Classification of Electrocardiogram (ECG) Beats Using Wavelet Transform and SVM and PCA-SVM

Abstract

A method for the automatic classification of cardio beats from an electrocardiogram (ECG) is presented in
the paper. This beats classification is based on an analysis of QRS and DWT based feature extraction. The principal
component analysis (PCA) is used for parameter analysis and recognition of cardiac beats. These parameters are
calculated for beats with 4 types of classes (L, A, P and R) from ECG records retrieved from the MIT-BIH arrhythmia
database. Further SVM is applied as classifier for automatic detection of heart beats. Analysis of the different groups
shows the overall recognition performance was 96.43% with SVM and 97.75% with PCA-SVM.]