Further properties of mixture designs for five components in orthogonal blocks
Orthogonally blocked experimental designs for mixtures of five ingredients, formed from Latin squares, were previously discussed by Prescott et al. Here, we extend this development by studying the properties of three classes of possible designs, with recommendations on their practical application. Restrictions on the design classes are explored and D-optimal (within the classes) versions are identified. Remarks on general D-optimality conclude the paper.
Year of publication: |
1997
|
---|---|
Authors: | Prescott, P. ; Draper, N. R. ; Lewis, S. M. ; Dean, A. M. |
Published in: |
Journal of Applied Statistics. - Taylor & Francis Journals, ISSN 0266-4763. - Vol. 24.1997, 2, p. 147-156
|
Publisher: |
Taylor & Francis Journals |
Saved in:
Saved in favorites
Similar items by person
-
Comparison of group screening strategies for factorial experiments
Dean, A. M., (2002)
-
Detection of interactions in experiments on large numbers of factors
Lewis, S. M., (2001)
-
On Rohlf's Method for the Detection of Outliers in Multivariate Data
Caroni, C., (1995)
- More ...