Exploring the Robustness of Wilcoxon Signed Rank-based Control Charts when symmetry is mis-specified Academic Article in Scopus uri icon

abstract

  • The Wilcoxon signed rank statistic is commonly used to build nonparametric control charts due to the unsensitivity of the statistic to deviations from the assumptions of normality. Initially used to test for symmetry, once symmetry can be assumed, it can also be employed to test for location changes. However, the performance of SR-based control charts may be severely affected when symmetry is incorrectly assumed at an early stage of analysis. This research explores the robustness and the effects of asymmetrical distributions and symmetry misspecification on the performances of a SR control chart for process monitoring. The results of this line of research will not only contribute to this area of research, but will assist working industrial and systems engineers make informed decision regarding the selection of nonparametric control charts in cases where symmetry is neither known or cannot be clearly identified. © IISE Annual Conference and Expo 2024.All rights reserved.

publication date

  • January 1, 2024