|
Avtomatika i Telemekhanika, 2013, Issue 10, Pages 55–67
(Mi at6148)
|
|
|
|
This article is cited in 14 scientific papers (total in 14 papers)
Topical issue
Properties of the posterior distribution of a regression model based on Gaussian random fields
A. A. Zaytsevab, E. V. Burnaevbac, V. G. Spokoinydc a Institute for Information Transmission Problems (Kharkevich Institute), Russian Academy of Sciences, Moscow, Russia
b SJC "Datadvance", Moscow, Russia
c Moscow Institute of Physics and Technology (State University), Dolgoprudny, Russia
d Weierstrass Institute, Berlin, Germany
Abstract:
We consider the regression problem based on Gaussian processes. We assume that the prior distribution on the vector of parameters in the corresponding model of the covariance function is non-informative. Under this assumption, we prove the Bernstein–von Mises theorem that states that the posterior distribution on the parameters vector is close to the corresponding normal distribution. We show results of numerical experiments that indicate that our results apply in practically important cases.
Citation:
A. A. Zaytsev, E. V. Burnaev, V. G. Spokoiny, “Properties of the posterior distribution of a regression model based on Gaussian random fields”, Avtomat. i Telemekh., 2013, no. 10, 55–67; Autom. Remote Control, 74:10 (2013), 1645–1655
Linking options:
https://www.mathnet.ru/eng/at6148 https://www.mathnet.ru/eng/at/y2013/i10/p55
|
Statistics & downloads: |
Abstract page: | 341 | Full-text PDF : | 212 | References: | 51 | First page: | 31 |
|