Analysis of the 12 run Plackett-Burman design (1997) [3 citations — 3 self]
Popular Tags
No tags have been applied to this document.
BibTeX | Add To MetaCart
author = {John Tyssedal and Oddgeir Samset},
title = {Analysis of the 12 run Plackett-Burman design},
institution = {},
year = {1997}
}
Bookmarks
OpenURL
Abstract:
Plackett-Burman designs whose run sizes are not a power of two, have lately received a lot of attention. Traditionally these designs have been recommended as main effects plans, since their complex alias pattern may create confusing results when standard methods of analysis are applied. This paper is concerned with the 12 run Plackett-Burman design. We present a new approach to the analysis of this design where both performance and implementation of methods are discussed. Under the assumption of effect sparsity we demonstrate how its very structured alias pattern leads to a new interpretation of halfnormal plots, and how the good projection properties may be exploited in searching for active factors. The suggested procedures seem to overcome some of the shortcomings reported by existing methods and make the 12 run PB design a most valuable screening design also when interactions are present. Key words: Alias structure; Design projectivity; Effect sparsity; Factor screening. 1 Introduct...
