基于LS-SVM残差控制图的失控信号诊断

The Diagnosis of the out-of-control Signals Based on LS-SVM Residual Chart

针对多变量统计过程控制中失控信号的诊断问题,提出了一种基于最小二乘支持向量机(LS-SVM)残差控制图的诊断方法,给出了基于LS-SVM残差控制图的构造方法与诊断程序,进行了实例分析与模拟分析,并与基于主成分分析的诊断方法进行了对比。研究结果表明:基于LS-SVM残差控制图的方法可以准确地诊断出失控变量,同时在小样本情况下依然适用。

In order to interpret the out of control signals in multivariate statistical process control,it proposes a method based LS-SVM residual chart,shows the detail about the method to develop LS-SVM residual chart,the diagnosis procedure,analysis of instance and simulation.It compares the method of LS-SVM residual chart with the methods of principal component analysis.The result shows that method of LS-SVM residual chart can diagnose out-of-control variables and correctly be used in small-size samples case.