基于SVM的步态信号模式分类研究

摘要

In order to identify the stride signal for patients with parkinson's disease,it proposes a method for extracting characteristic parameters from the stride pressure signal,establishes the SVM model of automatic stride pressure signal classfication,designs the classification process tool.The experiment shows that the method can identify more that 90 percent and be useful for detection and diagnosis of parkinson's disease patients%为了有效识别健康人和帕金森患者的步态信号,从步态压力信号入手,提出了一种有效提取特征参数的方法.建立了具有自动识别步态压力特征的SVM模型,设计出了支持向量机(SVM)的分类器.实验结果表明,该算法的识别率达到了90％以上,对帕金森患者的发现和诊断等具有一定的意义.