Finite Sample Bernstein – von Mises Theorem for Semiparametric Problems
Open Access
- 1 September 2015
- journal article
- Published by Institute of Mathematical Statistics in Bayesian Analysis
- Vol. 10 (3) , 665-710
Abstract
The classical parametric and semiparametric Bernstein – von Mises (BvM) results are reconsidered in a non-classical setup allowing finite samples and model misspecification. In the case of a finite dimensional nuisance parameter we obtain an upper bound on the error of Gaussian approximation of the posterior distribution for the target parameter which is explicit in the dimension of the nuisance and target parameters. This helps to identify the so called critical dimension $p_{n}$ of the full parameter for which the BvM result is applicable. In the important i.i.d. case, we show that the condition “$p_{n}^{3}/n$ is small” is sufficient for the BvM result to be valid under general assumptions on the model. We also provide an example of a model with the phase transition effect: the statement of the BvM theorem fails when the dimension $p_{n}$ approaches $n^{1/3}$. The results are extended to the case of infinite dimensional parameters with the nuisance parameter from a Sobolev class.
Keywords
All Related Versions
This publication has 28 references indexed in Scilit:
- Parametric estimation. Finite sample theoryThe Annals of Statistics, 2012
- The Bernstein-Von-Mises theorem under misspecificationElectronic Journal of Statistics, 2012
- On the Bernstein-von Mises phenomenon in the Gaussian white noise modelElectronic Journal of Statistics, 2011
- A high-dimensional Wilks phenomenonProbability Theory and Related Fields, 2010
- General frequentist properties of the posterior profile distributionThe Annals of Statistics, 2008
- Propagation-Separation Approach for Local Likelihood EstimationProbability Theory and Related Fields, 2005
- An MCMC approach to classical estimationJournal of Econometrics, 2003
- Asymptotic Normality of Semiparametric and Nonparametric Posterior DistributionsJournal of the American Statistical Association, 2002
- Wald Lecture: On the Bernstein-von Mises theorem with infinite-dimensional parametersThe Annals of Statistics, 1999
- Asymptotic Normality of Posterior Distributions in High-Dimensional Linear ModelsBernoulli, 1999