Reference : Nonparametric control charts: economic statistical design
E-prints/Working papers : First made available on ORBi
Business & economic sciences : Quantitative methods in economics & management
Nonparametric control charts: economic statistical design
Marcos Alvarez, Alejandro mailto [Université de Liège > Dép. d'électric., électron. et informat. (Inst.Montefiore) > Systèmes et modélisation >]
Heuchenne, Cédric mailto [Université de Liège > HEC-Ecole de gestion : UER > Statistique appliquée à la gestion et à l'économie >]
Faraz, Alireza mailto [Université de Liège > HEC-Ecole de gestion : UER > Statistique appliquée à la gestion et à l'économie >]
[en] Control charts ; Economic statistical design ; Nonparametric statistics ; Quality control ; Sign test ; Wilcoxon signed-rank test
[en] This paper studies economic statistical designs (ESD) for nonparametric control charts based on the sign and Wilcoxon tests. The main advantage of the procedures is that, except for the tested location parameter, they do not use either any parametric distribution for the quality characteristic or any information about the possible involved parameters, neither in the in-control nor in the out-of-control state. This is made possible by minimizing a cost function specified independently of these quantities. Unlike the ESD for the $\overline{x}$ chart, the resulting charts designs are robust to changes of the distributions of the observations (in control or out of control), provide reliable statistical guarantees when the $\overline{x}$ chart ESD does not and stay competitive even when the strong assumptions of the $\overline{x}$ chart ESD are fully satisfied. These new techniques can therefore be applied to a definitely wider class of problems and their designs may stay constant over time without losing performance.

File(s) associated to this reference

Fulltext file(s):

Open access
ijpe_2.pdfAuthor preprint433.24 kBView/Open

Bookmark and Share SFX Query

All documents in ORBi are protected by a user license.