脉搏信号和主成分分析在亚健康状态识别中的应用

摘要

为了评估亚健康状态,提出一种基于脉搏信号的亚健康状态识别新方法.用小波变换对脉搏信号消噪处理,再用功率谱、近似熵、小波熵估计提取特征量,对提取的特征量进行主成分分析,最后用改进的线性判别式分析法分类识别,主成分识别率达100％.该方法计算简单,稳定性好,识别率高,对亚健康状态的评估有一定的可行性.%In order to evaluate sub-health state, a novel sub-health state recognition method based on pulse signal is presented. Firstly, the wavelet transform is used to de-noise the pulse signal. Secondly, the power spectrum, approximate entropy and wavelet entropy estimation are employed to extract the characteristic quantity, and then to make principal component analysis on it. Finally, an improved linear discriminant analysis (LDA) is applied to classify and recognise. The rate of the principal component recognition is up to 100%. This method is simple in calculation, good in stability, and has high recognition rate, it is feasible to certain extent for sub-health state evaluation.