中文摘要：

目的：应用冠心病患者中医脉象非线性动力学特征建立证候诊断模型，旨在评价非线性分析方法对中医脉象客观化研究的意义。方法：应用ZBOX一1型中医脉象数字化分析仪记录528例冠心病患者中医脉图，基于支持向量机算法分别应用脉象信号时域特征参数和RQA特征参数并结合问诊、望诊参数建立冠心病证候诊断模型。结果：当问诊参数和望诊参数联合脉象信号时域特征参数建立证候模型时，证候模型的平均识别准确率相对于联合脉象信号前的模型有所提高，达70．17％；当脉象信号的RQA特征参数替代其时域特征参数参与建立证候模型后，证候模型的平均识别准确率显著提高，达82．83％。结论：脉象信号非动力学特征参与建立冠心病证候诊断模型能更大程度地提高模型的平均识别准确率。

英文摘要：

Objective： To establish syndrome diagnosis model with pulse nonlinear dynamics characteristics of patients with coronary heart disease, to evaluate the significance of pulse nonlinear dynamic method. Methods： To recording the pulse graph of patients with coronary heart disease with ZBOX-1 pulse apparatus, establishing syndromes diagnosis model with pulse time-domain parameters and RQA parameters combining the inquiry, inspection parameters based on support vector machine algorithm. Results： The average recognition rate with pulse time-domain parameters and inquiry, inspection parameters has reached 70.17%, which compared with pulse time-domain parameters; the average recognition rate with RQA parameters and inquiry, inspection parameters has increased to 82.83%. Conclusion： The average recognition rate of syndromes diagnosis model will increase with pulse nonlinear dynamics characteristics. It provides a new idea and method for objective study of pulse in TCM.