B0669
Title: Change-point control charts in the presence of a tail-shift of the underlying distribution
Authors: Claudio Giovanni Borroni - University of Milano - Bicocca (Italy) [presenting]
Manuela Cazzaro - University of Milano-Bicocca (Italy)
Paola Maddalena Chiodini - University of Milano-Bicocca (Italy)
Abstract: The change-point methodology can be differently implemented to get some known and some new self-starting control charts. Non-parametric charts are considered based on the Cramer-Von-Mises test for the equality of two completely unspecified distribution functions. A simulation study is reported to understand whether the choice of implementation can affect the chart's performance in terms of the average number of further readings needed to get a signal after a shift of the underlying distribution occurs. Sever kinds of shift can be considered: beyond the trivial case of a shift in mean, the focus is on shifts in variability and on their relationship with shifts in the whole shape of the distribution, especially in the tails. Some insight about the best implementation when the tails are burdened or inflated is obtained, and it is attempted to link those shifts to real-life situations encountered in some production processes.