An Electrocardiogram Classification Method Combining Morphology Features

In this paper, an expert experience based Electrocardiogram (ECG) classification method using domain knowledge and morphology information is presented. Firstly, the process of ECG interpretation by physicians is analyzed. Then, the construction method of classification model based on Support Vector Machine (SVM) is discussed and morphology information extraction approach through Principal Component Analysis and Independent Component Analysis is emphasized. Finally, entropy is introduced to evaluate the effectiveness of different feature spaces for abnormal ECG detection. Totally 94325 heart beats from MIT-BIH Arrhythmia Database and 289 12-lead records from Chinese Cardiovascular Disease Database are used to verify the classification model respectively. According to experiment results, the accuracy of classifier is improved.