On asymptotic distribution of optimal design for polynomial-type regression
D-optimal designs [xi]*k,s for a class of polynomial-type regression model tfk,s(x) = vb x vb5(x, x2,..., xk), s> -1, on the intervals [a, 1], 0 [less-than-or-equals, slant] a < 1, considered. The D-optimal designs [xi]*k,s as k --> [infinity], converge weakly to the arcsin distribution on [a, 1]. Comparisons of [xi]*k,s with the arcsin distribution [xi][infinity] are also made.
| Year of publication: |
1998
|
|---|---|
| Authors: | Chang, Fu-Chuen |
| Published in: |
Statistics & Probability Letters. - Elsevier, ISSN 0167-7152. - Vol. 36.1998, 4, p. 421-425
|
| Publisher: |
Elsevier |
| Keywords: | Arcsin distribution D-efficiency D-optimal design Jacobi polynomial Polynomial-type |
Saved in:
Saved in favorites
Similar items by person
-
On Minimally-supported D-optimal Designs for Polynomial Regression with Log-concave Weight Function
Chang, Fu-Chuen, (2007)
-
Exact D-optimal designs for polynomial regression without intercept
Chang, Fu-Chuen, (1999)
-
D-optimal designs for polynomial regression with exponential weight function
Chang, Fu-Chuen, (2009)
- More ...