基于混沌参数和改进的重现定量分析的心电诊断

Diagnosis of ECG Signal Based on Chaos Parameters and Improved Recurrence Quantification Analysis

目的:对正常窦性心律(SNR)、室速(VT)和室颤(VF)3种心电信号进行分类。方法:将非线性动力学参数与重现定量分析(RQA)相结合,选取参数组成多重特征向量。非线性动力学参数选取关联维和最大李雅普诺夫指数,重现定量分析参数选取重现率,结合心电信号特征提出参数L/V。结果:使用lib-svm对这3类心电信号进行分类,得到了较高的正确率。结论:该算法能够有效地对心电信号进行分类,可用于心电的自动化诊断。

Objective To classify the three kinds of ECG signals,including normal sinus,ventricular tachycardia and ventricular fibrillation.Methods Nonlinear dynamics parameters and recurrence quantification analysis were combined to form feature vector.In the nonlinear dynamics parameters,maximal Lyapunov exponent and correlation dimension were chosen while in recurrence quantification,recurrence rate was selected and parameter L/V was proposed according to the characteristic of ECG signals.Results Accuracy was satisfied by using lib-svm to classify the three kinds of ECG signals.Conclusion The Algorithm can classify ECG signals efficiently and can be used in automatic diagnosis of ECG signal.