Support Vector Machine-based ECG Beats Automatic Diagnosis System

摘要

Correct recognition and effective classification of ECG heartbeats are fundamental and essential for clinical therapy. The paper proposes a system, consisting of three procedures, raw signal preprocessing, feature extraction and the core part classification, for automatic diagnosis of heartbeats. The multi-class support vector machine with Q-Gaussian kernel (QG-MSVM) is employed for classifier. With the standard MIT/BIH arrhythmia database, simulation experiments are carried out based on LIBSVM tool. Experimental results indicate the proposed method is better than other researchers' counterparts.