Detection of ventricular fibrillation rhythm by using boosted support vector machine with an optimal variable combination

Abstract

In this paper, the ventricular fibrillation (VF) rhythm is detected by using a new approach involving the support vector machine (SVM), adaptive boosting (AdaBoost) and differential evolution (DE) algorithms with the help of an optimal variable combination. The proposed methodology has been validated on training sets and testing sets that were obtained from three databases, namely MIT-BIH malignant ventricular arrhythmia database, arrhythmia database, and CUDB database. In the evaluation phase, the proposed methodology shows superior performance in detection of the VF rhythm than competing methods: an accuracy of 98.20%, a sensitivity of 98.25%, and specificity of 98.18% using 5 s of the ECG segments. Another advantage of our method is that it needs less memory and can be implemented in real-time.