Multivariate measurement system analysis in multisite testing: An online technique using principal component analysis

Determining the optimal allocation of parameters for multivariate measurement system analysis

“…Other multivariate measurement system analysis methods using PCA related methods can be referred to He et al (2011), Osma (2011 and Peruchi et al (2013Peruchi et al ( , 2014. Although the PCA method can be used to transform the correlated characteristics into independent variables and these independent variables are analyzed using a random factor experiment, these methods cannot provide/remove the causes of production problems.…”

“…Recently, He et al (2011) proposed a modified principal component analysis (PCA) method to transform multivariate measurement data with dependent variables into those with independent principal components. Although He's and the other existing PCA methods can be served as an alternative approach to multivariate MSA, these methods cannot provide and remove the causes of production problems.…”

“…Recently, principle component analysis (PCA) related methods are proposed by He, Wang, and Cook (2011), Osma (2011), Peruchi, Balestrassi, de Paiva, Ferreira, and de Santana Carmelossi (2013) and Wang (2013) for analyzing gauge variation. He et al (2011) proposed a modified principal component analysis (PCA) method to transform multivariate measurement data with dependent variables into those with independent principal components.…”

“…He et al (2011) proposed a modified principal component analysis (PCA) method to transform multivariate measurement data with dependent variables into those with independent principal components. Peruchi et al (2013) and Peruchi, Paiva, Balestrassi, Ferreira, and Sawhney (2014) proposed a weighted principal component (WPA) method for multivariate analysis of a measurement system, which can be considered as a special PCA approach.…”

On the assessment of multivariate and multisite measurement systems

“…Multisite measurement systems (He et al . ) are the typical cases where there are several measurement devices. A multisite measurement system has u ( u > 1) measurement instruments in parallel where each instrument independently measures multivariate quality characteristics: a unit coming from the manufacturing process is randomly assigned to one of the measurement instruments, then once measured the units are merged back together before entering in a succeeding step of the manufacturing process.…”