Abstract
Commutative Jordan algebras are used to drive an highly tractableframework for balanced factorial designs with a prime number p oflevels for their factors. Both fixed effects and random effects modelsare treated. Sufficient complete statistics are obtained and used toderive UMVUE for the relevant parameters. Confidence regions areobtained and it is shown how to use duality for hypothesis testing.
Original language | English |
---|---|
Pages (from-to) | 15-25 |
Number of pages | 11 |
Journal | Discussiones Mathematicae: Probability and Statistics |
Volume | 27 |
Publication status | Published - 1 Jan 2007 |
Keywords
- prime basis factorial
- commutative Jordan algebras
- complete sufficient statistics
- UMVUE
- confidence regions