von Rosen, Dietrich
- Institutionen för energi och teknik, Sveriges lantbruksuniversitet
Sammanfattning
The objective of this paper is to consider shift invariance of random factors in linear models. Marginally shift invariant interaction factors are treated. The random factors are described via their covariance matrices and it is shown that shift invariance implies Toeplitz covariance matrices and marginally shift invariance implies block Toeplitz covariance matrices. In order to get interpretable linear models reparameterization is taken place and it is shown that by putting restrictions on the spectrum of the Toeplitz matrices natural reparameterization conditions are obtained
Nyckelord
Block Toeplitz matrix; covariance structures; eigenspace; invariance; K-way tables; marginal permutations; reparameterization; shift invariance; spectrum; Toeplitz matrix
Publicerad i
Research report (Centre of Biostochastics)
2005, nummer: 2005:6
Utgivare: Centre of Biostochastics, Swedish University of Agricultural Sciences
SLU författare
UKÄ forskningsämne
Jordbruksvetenskap
Sannolikhetsteori och statistik
Permanent länk till denna sida (URI)
https://res.slu.se/id/publ/6849