Classification ECG of Cardiac Signals Using LPC Features and Support Vector Machine

Cardiovascular diseases are of the most common diseases in the world and one of the 3 main causes of death. If the Cardiovascular diseases are diagnosed and treated in the early stages, they are very effective in patient’s health. In this paper, diagnosis of cardiovascular disease type by Electrocardiogram signal (ECG) or by sound of PCG signal is done by using different features. Also, it is tried to use Features extracted from the ECG signal as a tool to develop therapy, research and diagnostic areas by using different protocols. In this study, Classifying 5 arrhythmia samples has been done from ECG cardiac signals with the LPC linear prediction coefficients features And SVM classification. For this purpose, each signal is framed to time intervals 1 to 5 seconds and for each frame, some LPC coefficients are calculated. The framing results with different interval 1 to 5 seconds were examined and observed by using LPC method, 1 second framing has better results. Also, extracted features results are compared with wavelet features, too. In the suggested method, we can obtain to accuracy higher than 99%.

Keywords:

ECG , Feature Extraction , LPC , PCG , SVM