Further Characterizations of Design Optimality and Admissibility for Partial Parameter Estimation in Linear Regression
Gaffke, Norbert
Ann. Statist., Tome 15 (1987) no. 1, p. 942-957 / Harvested from Project Euclid
Text on Project Euclid
PDF
Alt PDF
The paper gives a contribution to the problem of finding optimal linear regression designs, when only $s$ out of $k$ regression parameters are to be estimated. Also, a treatment of design admissibility for the parameters of interest is included. Previous results of Kiefer and Wolfowitz (1959), Karlin and Studden (1966) and Atwood (1969) are generalized. In particular, a connection to Tchebycheff-type approximation of $\mathbb{R}^s$-valued functions is found, which has been known in case $s = 1$. Strengthened versions of the results are obtained for invariant designs in situations, when invariance properties of the regression setup are available. Applications are given to multiple quadratic regression and to one-dimensional polynomial regression.
Publié le : 1987-09-14
Classification:  Gauss-Markov estimator,  approximate design theory,  optimality criterion,  equivalence theorem,  minimax theorem,  Tchebycheff approximation,  invariant design,  multiple quadratic regression,  polynomial regression,  62K05,  49B40 
@article{1176350485,
     author = {Gaffke, Norbert},
     title = {Further Characterizations of Design Optimality and Admissibility for Partial Parameter Estimation in Linear Regression},
     journal = {Ann. Statist.},
     volume = {15},
     number = {1},
     year = {1987},
     pages = { 942-957},
     language = {en},
     url = {http://dml.mathdoc.fr/item/1176350485}
}

Gaffke, Norbert. Further Characterizations of Design Optimality and Admissibility for Partial Parameter Estimation in Linear Regression. Ann. Statist., Tome 15 (1987) no. 1, pp.  942-957. http://gdmltest.u-ga.fr/item/1176350485/