Spacing of information in polynomial regression:a simple solution
Consider a k polynomial regression on a single real variable. If n uncorrelated observations are to be taken in a design with support on more than \(k+1\) points, there is an approximate experiment, \(\nu\), with support on \(k+1\) points and n observations such that both designs have the same information matrix for the model. A proof of this result is provided. A method to obtain the approximate design \(\nu\) is given and illustrated by an example. The source of disagreement between \textit{J. Kiefer} [J. R. Stat. Soc., Ser. B 21, 272-319 (1959; Zbl 0108.153)] and \textit{A. de la Garza} [Ann. Math. Stat. 25, 123-130 (1954; Zbl 0055.132)] in the solution of this problem is clarified.
- Equally spaced design points in polynomial regression: A comparison of systematic sampling methods with the optimal design of experiments
- scientific article; zbMATH DE number 4072184
- scientific article; zbMATH DE number 1209408
- scientific article; zbMATH DE number 1747150
- Approximate optimal designs for multivariate polynomial regression
This page was built for publication: Spacing of information in polynomial regression:a simple solution
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3678501)