因子篩選實驗中活性因子辨識力的穩健性探討

A Robustness Study for the Identification of Active Factors in Screening Experiments

俞一唐(I-Tang Yu)；張森傑(Sen-Chieh Chang)

交叉驗證法 ； 改良後Box-Meyer法 ； 因子篩選實驗 ； Cross-validation ； Modified Box-Meyer method ； Screening experiment

中國統計學報

61卷3期（2023 / 09 / 01）

231 - 248

繁體中文;英文

因子篩選實驗是在一個正式實驗前，用來從大量的潛在因子中篩選出重要活性因子的前期實驗。改良後Box-Meyer法（modified Box-Meyer method, MBMM）是將Box-Meyer法中，各候選模型之下所使用的效果模型以平均數模型取代。因此，無論在候選模型的建構與概似函數的假設方面，MBMM皆以因子為考量，所以非常適用於因子篩選實驗資料的分析。本文分別利用交叉驗證法與統計模擬來評估MBMM所辨識出活性因子的穩健性與正確性。在我們分析的兩個實例中，上述兩種方式皆顯示MBMM對於活性因子的篩選非常具有穩健性與正確性。

Before running a large-scale experiment, screening experiments can be constructed to identify active factors among a large number of potential factors. Instead of an effects model used in the Box-Meyer method, the modified Box-Meyer method (MBMM) assumes a means model in each conditional model. As a result, the MBMM is based only on the factor combinations for both the construction of the candidate models and the specification of the likelihood functions, which makes the MBMM suitable for analyzing screening experiments. In this paper, we use both the cross-validation and simulation to investigate the robustness of the MBMM. From the results of analyzing two examples, we conclude that the MBMM is very robust in the identification of active factors.

1. Box, G. E. P.,Meyer, R. D.(1993).Finding the active factors in fractionated screening experiments.Journal of Quality Technology,25(2),94-105.
2. Hoeting, J. A.,Madigan, D.,Raftery, A. E.,Volinsky, C. T.(1999).Bayesian model averaging: A tutorial.Statistical Science,14(4),382-417.
3. James, G.,Witten, D.,Hastie, T.,Tibshirani, R.(2013).An Introduction to Statistical Learning.United States:Springer.
4. Jang, D. H.,Anderson-Cook, C. M.(2018).Assessing robustness of factor ranking for supersaturated designs.Quality and Reliability Engineering International,34(3),417-426.
5. Montgomery, D. C.(2019).Design and Analysis of Experiments.United States:John Wiley & Sons.
6. Raftery, A. E.,Madigan, D.,Hoeting, J. A.(1997).Bayesian model averaging for linear regression models.Journal of the American Statistical Association,92(437),179-191.
7. Wu, C. F. J.,Hamada, M.(2009).Experiments: Planning, Analysis, and Parameter Design Optimization.United States:John Wiley & Sons.
8. Yu, I. T.(2016).A Bayesian approach to the identification of active location and dispersion factors.IIE Transactions,48(7),629-637.
9. Yu, I. T.(2013).A modification of the Box–Meyer method for finding the active factors in screening experiments.Journal of Applied Statistics,40(5),972-984.
10. Yu, I. T.(2022).A Bayesian approach to factor screening in life tests.Communication in Statistics- Theory and Methods,51(6),1778-1790.