Frequentist Model Average Estimators
Nils Lid Hjort, Gerda Claeskens · Journal of the American Statistical Association · 2003
The traditional use of model selection methods in practice is to proceed as if the final selected model had been chosen in advance, without acknowledging the additional uncertainty introduced by model selection. This often means underreporting of variability and too optimistic confidence intervals. We build a general large-sample likelihood apparatus in which limiting distributions and risk properties of estimators post-selection as well as of model average estimators are precisely described, also explicitly taking modeling bias into account. This allows a drastic reduction in complexity, as competing model averaging schemes may be developed, discussed, and compared inside a statistical prototype experiment where only a few crucial quantities matter. In particular, we offer a frequentist view on Bayesian model averaging methods and give a link to generalized ridge estimators. Our work also leads to new model selection criteria. The methods are illustrated with real data applications.
Cited by 688 works. doi:10.1198/016214503000000828
Authors
References
- Applied logistic regression
- Estimating the Dimension of a Model
- Theory of Point Estimation
- Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach
- Approximation Theorems of Mathematical Statistics
- Discussion on the paper by Spiegelhalter, Best, Carlin and van der Linde
- Regression and Time Series Model Selection in Small Samples
- Statistical Modeling: The Two Cultures (with comments and a rejoinder by the author)
- Bayesian model averaging: a tutorial (with comments by M. Clyde, David Draper and E. I. George, and a rejoinder by the authors
- Assessment and Propagation of Model Uncertainty