Resumen de Self-Starting Multivariate Control Charts for Location and Scale
Edgard M. Maboudou-Tchao, Douglas M. Hawkins
The multivariate exponentially weighted moving average and multivariate exponentially weighted moving covariance matrix are useful control charts for monitoring correlated characteristics, but they assume the parameters are already known. In practice this is rarely the case, and estimates are typically used. This fundamentally changes the results and leads to poor chart performance. Self-starting methods provide the necessary initial parameters. Applying self-starting methods to moving average and moving covariance control charts is here shown to produce good results.
