Automatic Classification of ECG Arrhythmia Using Morphological Parameters with HMM and SVM

A method for the automatic classification of cardiopathies from an electrocardiogram (ECG) is presented in the paper. This treatment is based on an analysis of certain morphological parameters for the recognition of 4cardiopathies. The Hidden Markov Model (HMM) was used for parameter analysis and recognition of cardiac arrhythmias. The morphological parameters were divided into homogeneous groups (amplitude, surface, interval and slope). These parameters are calculated for beats with 4 types of abnormalities (RBBB, APC, PVC and LBBB) from ECG records retrieved from the MIT-BIH arrhythmia database. Further SVM is applied as classifier for automatic detection of heart disease. Analysis of the different groups shows the overall recognition performance was 98.43%. The worst is 96.75% for the RBBB class.