EXPLORATION OF BIAS IN EXPERIMENTAL DESIGN

探討模型偏誤於實驗設計階段相關影響之研究

曾世賢(Shih-Hsien Tseng) ； 陳曉瑩(Hsiao-Yin Chen)

《品質學報》 24卷2期 (2017/04) Pp. 92-111

https://doi.org/10.6220/joq.2017.24(2).02

摘要

In 1959, George Box and other scientists became the first to study the effect of model misspecification errors or ＂bias＂ in experimental design. They found that bias has a far greater effect on accuracy than variance and that efforts to mitigate the effects of bias generally help with other errors, but not the reverse. The propose of this paper is to discuss the research attempts to mitigate the effects of bias at the planning stage of experimentation and what assumption methods are relevant to a new generation of planning and analytic techniques designed to mitigate these effects. Thus, we will review the definition of bias, assumptions about it, its effect, and several criteria for determining its presence that are used to obtain optimal designs and evaluate bias sensitivity, including irregularly shaped design regions.

關鍵字

bias ； variance ； design of experiment ； Bayesian ； optimal designs

並列摘要

1959年，George Box與其他學者首先探討實驗設計中模型偏誤所造成之影響。根據發現，模型偏誤對預測準確度所造成之影響更甚於變異數且當我們嘗試降低偏誤時也可間接減少變異數所造成之影響，反之則不然。本篇論文主要討論在實驗設計階段相關學者所提出降低偏誤之方法、相關之假設與分析之手法等。因此我們回顧之前有關研究，包含偏誤之定義、假設、影響與取得最佳化設計之有關準則，此外本篇論文也提供關於偏誤對於模型準確度在不規則形狀之設計區之敏感度相關討論。

並列關鍵字

偏誤 ； 變異 ； 實驗設計 ； 貝氏估計 ； 最佳化設計

參考文獻

Allen, T. T.(2006).Introduction to Engineering Statistics and Six Sigma.London:Springer.

Allen, T. T.,Tseng, S. H.(2011).Variance plus bias optimal response surface designs with qualitative factors applied to stem choice modeling.Quality and Reliability Engineering International.27(8),1199-1210.

Allen, T. T.,Yu, L.(2002).Low-cost response surface methods from simulation optimization.Quality and Reliability Engineering International.18(1),5-17.

Allen, T. T.,Yu, L.,Schmitz, J.(2003).An experimental design criterion for minimizing meta-model prediction errors applied to die casting process design.Journal of the Royal Statistical Society Series C: Applied Statistics.52(1),1-15.

Beattie, S. D.,Fong, D. K. H.,Lin, D. K. J.(2002).A two-stage Bayesian model selection strategy for supersaturated designs.Technometrics.44(1),55-63.