Optimal designs for frequentist model averaging
| Autor: | Kirsten Schorning, Kira Alhorn, Holger Dette |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
Statistics and Probability
Optimal design FOS: Computer and information sciences Optimality criterion Mean squared error General Mathematics Bayesian probability 01 natural sciences Methodology (stat.ME) 010104 statistics & probability Frequentist inference 0502 economics and business Applied mathematics 050207 economics 0101 mathematics Statistics - Methodology Mathematics Applied Mathematics Model selection 05 social sciences Estimator Regression analysis Articles Agricultural and Biological Sciences (miscellaneous) Statistics Probability and Uncertainty General Agricultural and Biological Sciences |
| Zdroj: | Biometrika |
| DOI: | 10.48550/arxiv.1807.05234 |
| Popis: | SummaryWe consider the problem of designing experiments for estimating a target parameter in regression analysis when there is uncertainty about the parametric form of the regression function. A new optimality criterion is proposed that chooses the experimental design to minimize the asymptotic mean squared error of the frequentist model averaging estimate. Necessary conditions for the optimal solution of a locally and Bayesian optimal design problem are established. The results are illustrated in several examples, and it is demonstrated that Bayesian optimal designs can yield a reduction of the mean squared error of the model averaging estimator by up to 45%. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|