基于WD-LSSVM的监护信息预报研究  

Study on Prediction of Monitoring Information Based on WD-LSSVM

作　　者：蒋贤海[1] 谢存禧[2] 

机构地区：[1]广东水利电力职业技术学院机械工程系,广东广州501635 [2]华南理工大学机械与汽车工程学院,广东广州510640

出　　处：《机电产品开发与创新》2013年第2期112-114,共3页Development & Innovation of Machinery & Electrical Products

基　　金：粤港关键领域重点突破项目(20090101-1)

摘　　要：针对监护信息预报问题,提出了一种新的监护信息预报方法。采用小波分解和最小二乘支持向量机相结合的方法,该方法利用小波分解将原参数序列分解为逼近信号和细节信号,采用LS-SVM算法,分别建立监护信息逼近信号和细节信号的预报模型,得到逼近信号和细节信号的监护信息预报值,然后将各部分预报值利用小波逆变换进而得到监护信息最终预报结果。最后对提出的方法进行了实验验证。According to the forecast problem of monitoring information, the method of wavelet decomposition and LS-SVM is put forward. Physiological parameter is decomposed into approximation signal and the detail signal by means of the method, and the forecast results are obtained with prediction models of the approximation signal and the detail signal which are established respectively using LS-SVM, and then the final prediction results of monitoring information using inverse wavelet transform with the forecast results before. Finally the proposed method is verified by experiment.

关 键 词：监护信息 小波分解 最小二乘支持向量机 预报 

分 类 号：TP23[自动化与计算机技术—检测技术与自动化装置]