This is a toolkit for multivariate process monitoring. It computes several multivariate control charts e.g. Hotelling, Chi-squared, MEWMA, MCUSUM and Generalized Variance. Ten didactic datasets are included. It includes some techniques for assessing multivariate normality e.g. Mardia's, Royston's and Henze-Zirkler's tests. Please, see the NEWS file for the latest changes in the package.
Changes in the Version 1.0.2 from 2016-05-30.
LINE 71 mult.chart
The UCL for individual observations in Phase I is : ucl <- ((colm - 1) ^ 2) / colm * qbeta(1 - alpha, p / 2, ((colm - p - 1) / 2) ) See Tracy, Young, and Mason (1992) “Multivariate Control Charts for Individual Observations,”
The previous UCL was ucl <- ((colm - 1) ^ 2) / colm * qbeta(1 - alpha, p / 2, (((2 * (colm - 1) ^ 2) / (3 * colm - 4) - p - 1) / 2)), which was propossed by Sullivan and Woodall (1996). “A Comparison of Multivariate Control Charts for Individual Observations,” This method is used by MINITAB 17.