Skip to Main content Skip to Navigation
Journal articles

Automatic generation of generalised regular factorial designs

Abstract : The R package planor enables the user to search for, and construct, factorial designs satisfying given conditions. The user specifies the factors and their numbers of levels, the factorial terms which are assumed to be non-zero, and the subset of those which are to be estimated. Both block and treatment factors can be allowed for, and they may have either fixed or random effects, as well as hierarchy relationships. The designs are generalised regular designs, which means that each one is constructed by using a design key and that the underlying theory is that of finite abelian groups. The main theoretical results and algorithms on which planor is based are developed and illustrated, with the emphasis on mathematical rather than programming details. Sections 3-5 are dedicated to the elementary case, when the numbers of levels of all factors are powers of the same prime. The ineligible factorial terms associated with users' specifications are defined and it is shown how they can be used to search for a design key by a backtrack algorithm. Then the results are extended to the case when different primes are involved, by making use of the Sylow decomposition of finite abelian groups. The proposed approach provides a unified framework for a wide range of factorial designs.
Document type :
Journal articles
Complete list of metadata

Cited literature [43 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01608689
Contributor : Archive Ouverte Prodinra <>
Submitted on : Monday, May 25, 2020 - 3:37:56 PM
Last modification on : Monday, April 12, 2021 - 1:42:24 PM

File

2017_Monod_Computational Stati...
Publisher files allowed on an open archive

Licence


Distributed under a Creative Commons Attribution - ShareAlike 4.0 International License

Identifiers

Collections

Citation

Andre Kobilinsky, Hervé Monod, R.A. Bailey. Automatic generation of generalised regular factorial designs. Computational Statistics and Data Analysis, Elsevier, 2017, 113, pp.311-329. ⟨10.1016/j.csda.2016.09.003⟩. ⟨hal-01608689⟩

Share

Metrics

Record views

158

Files downloads

61