A Study on Criterion-Robust Discrimination Designs
| Autor: | Yin-Shi Chang, 張殷實 |
|---|---|
| Rok vydání: | 2009 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 97 In this thesis, we tackle the problem of discriminating among polynomial regression models with degree k 、k-1 、k-2 on the interval [-1,1]. We attempt to construct some designs which are efficient for two objectives : (1) to discriminate between models with degree k and k-1, and (2) to discriminate between models with degree k-1 and k-2. For the dual objective we propose a multipleobjective optimality criterion --- Φ_β-optimality criterion. By the aid of canonical moments and maximin principle, we derive a criterion-robust discrimination design for the described problem. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|