A Bayesian approach to factor screening in life tests. Issue 6 (19th March 2022)

• Record Type: Journal Article
• Authors: Yu, I-Tang
• Is Part Of: Communications in statistics. Volume 51:Issue 6(2022)
• Page Start: 1778
• Page End: 1790
• Publication Date: 2022-03-19
• Subjects: Bayesian model averaging -- Box-Meyer method -- designed experiment -- Empirical Bayes estimation -- screening experiment; Mathematical statistics -- Periodicals; Mathematics; Statistics; 519.2
• DOI: 10.1080/03610926.2020.1768270
• Languages: English
• ISSNs: 0361-0926

Abstract

Abstract: The modified Box-Meyer method (MBMM) has been proposed to identify active factors in unreplicated screening experiments. This paper aims to introduce the MBMM into the analysis of screening experiments with lifetime data. Experiments both with and without replicates are considered. Censored observations arise commonly in lifetime data which increases computational complexity when applying the MBMM. Unlike the original MBMM, we propose a quasi-empirical Bayes approach to estimate the hyper-parameters. By doing so, the computational complexity is reduced. We illustrate the proposed approach by analyzing two well-studied examples, and all the active factors are identified successfully.