主成分分析法在多元质量控制中的应用  

Principle Component Analysis Application in Multivariate-quality-control

作　　者：王亚雄[1] 李建英[1] 

机构地区：[1]湖南大学统计学系,湖南长沙410079

出　　处：《工业工程与管理》2005年第3期121-125,共5页Industrial Engineering and Management

摘　　要：提出以多元统计方法中的主成分分析法解决大型企业中的多元质量控制问题。阐明了主成分分析法的原理,在观测变量为正态分布的假设下,论述了Q Q图、主成分质量椭圆图以及T2控制图的原理及构造方法。通过化妆品公司的实例说明了主成分分析法的两个作用:一是发现异常值,二是进行质量稳定性控制。The principle component analysis is proposed to solve the Multivariate-quality-control problem existed in some big enterprises, and the theory of principle component analysis is introduced. Under the assumption that observing variables are normal distribution, the methods introduced to draw the Q-Q、principle component quality oval diagram and T 2 quality control chart are discussed. The two functions of principle component analysis: one is to find the abnormal value and the other is to control quality stability are also described.

关 键 词：主成分分析法 多元质量控制 主成分质量椭圆 T^2质量控制 

分 类 号：F406.3[经济管理—产业经济]