Document ESAIM: P&S, March 2006, Vol. 10, pp. 164-183
DOI: 10.1051/ps:2006004
Model selection for estimating the non zero components of a Gaussian vector
Sylvie HuetINRA, MIA, 78352 Jouy-en-Josas Cedex, France; huet@banian.jouy.inra.fr
(Received January 13, 2004. Revised September 28, 2005. / Published online: 9 March 2006)
Abstract
We propose a method based on a penalised likelihood criterion, for
estimating the number on non-zero components of the mean
of a
Gaussian vector. Following the work of Birgé and Massart in Gaussian model
selection, we choose the penalty function such that the resulting
estimator minimises the Kullback risk.
Mathematics Subject Classification. 62G05, 62G09
Key words: Kullback risk, model selection, penalised likelihood criteria.
© EDP Sciences, SMAI 2006