Optimization of Barron density estimates

0410591 - UTIA-B 20010060 RIV US eng J - Článek v odborném periodiku
Vajda, Igor - van der Meulen, E. C.
Optimization of Barron density estimates.
IEEE Transactions on Information Theory. Roč. 47, č. 5 (2001), s. 1867-1883. ISSN 0018-9448. E-ISSN 1557-9654
Grant CEP: GA ČR GA102/99/1137
Grant ostatní: Copernicus(XE) 579
Výzkumný záměr: AV0Z1075907
Klíčová slova: Barron estimator * chi-square criterion * density estimation
Kód oboru RIV: BD - Teorie informace
Impakt faktor: 2.077, rok: 2001

We investigate a nonparametric estimator of probability density introduced by A.R.Barron. Earlier papers established its consistency in a strong sense,e.g.,in the expected information divergence or expected chi-square divergence. This paper pays main attention to the expected chi-square divergence criterion. We give a new motivation of the Barron estimators by showing that a maximum-likelihood estimator (MLE) of a density from a family important in practice is consistent in expected information divergence.
Trvalý link: http://hdl.handle.net/11104/0130680