The efficiency of two Bayesian order estimators is studied. By using nonparametric techniques, we prove new underestimation and overestimation bounds. The results apply to various models, including mixture models. In this case, the errors are shown to be O(e−an) and $O((\log n)^{b}/\sqrt{n})$ (a, b>0), respectively.
Publié le : 2008-04-15
Classification:
Mixture,
model selection,
nonparametric Bayesian inference,
order estimation,
rate of convergence,
62F05,
62F12,
62G05,
62G10
@article{1205420524,
author = {Chambaz, Antoine and Rousseau, Judith},
title = {Bounds for Bayesian order identification with application to mixtures},
journal = {Ann. Statist.},
volume = {36},
number = {1},
year = {2008},
pages = { 938-962},
language = {en},
url = {http://dml.mathdoc.fr/item/1205420524}
}
Chambaz, Antoine; Rousseau, Judith. Bounds for Bayesian order identification with application to mixtures. Ann. Statist., Tome 36 (2008) no. 1, pp. 938-962. http://gdmltest.u-ga.fr/item/1205420524/