Acta Oeconomica Pragensia 2007, 15(1):27-35 | DOI: 10.18267/j.aop.35
Analysis of Dispersion Effects for Robust Design
- doc. Ing. Eva Jarošová, CSc.; Katedra statistiky a pravděpodobnosti, Fakulta informatiky a statistiky, VŠE v Praze, jarosova@vse.cz.
- Ing. Pavel Zimmermann - doktorand; Katedra statistiky a pravděpodobnosti, VŠE v Praze, zimmerp@vse.cz.
Robust design reduces excessive process variation. Identification of dispersion effects plays an important role in it. Two main approaches are used to analyze the experiment where levels of noise factors are changed on purpose. Either the function of sample variance computed from replications over various levels of the noise factors, or response values themselves are modelled. The presence of heteroscedasticity may complicate the identification of dispersion effects. The aim of the simulation study described in the paper was to compare various ways of modelling heteroscedasticity and the efficiency of both main approaches. We confined ourselves to the assumption of normal random errors and used the t-test to identify the dispersion effects.
Keywords: dispersion effects, heteroscedasticity, generalized linear model
JEL classification: G30
Published: February 1, 2007 Show citation
References
- DAVIDIAN, M. - CARROLL, R. J., 1987: Variance Function Estimation. JASA, Theory and Methods, 1987, roč. 82, č. 400, s. 1079-1091.
Go to original source...
- PIGNATIELLO, J. J. - RAMBERG, J. S., 1985: "Discussion of 'Off-Line Quality Control, Parameter Design, and the Taguchi Method' by Kackar, R.N.," Journal of Quality Technology, 1985, roč. 17, s. 198-206.
Go to original source...
- McCULLAGH, P. - NELDER, J. A., 1989: Generalized Linear Models. London, Chapman & Hall, 1989.
Go to original source...
- WU, C. F. J. - HAMADA, M., 2000: Experiments. Planning, Analysis, and Parameter Design Optimization. New York, J. Wiley & Sons, 2000.
This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 International License (CC BY 4.0), which permits use, distribution, and reproduction in any medium, provided the original publication is properly cited. No use, distribution or reproduction is permitted which does not comply with these terms.