no. 11-12
Statistics
Unbiased risk estimation and scoring rules
[Estimation de risque non biaisée et règles dʼévaluation]

Présenté par : Paul Deheuvels

Ehm, Werner1
1 Institute for Frontier Areas of Psychology and Mental Health, Wilhelmstr. 3a, 79098 Freiburg, Germany
Comptes Rendus. Mathématique, Tome 349 (2011) no. 11-12, pp. 699-702.

La méthode dʼestimation du risque Steinien est doublement généralisée, dʼune part du modèle de translation Gaussien à des familles non paramétriques et dʼautre part du risque quadratique à des distances du type divergence plus générales. Cette extension repose sur une relation avec des règles dʼévaluation locales propres.

Stein unbiased risk estimation is generalized twice, from the Gaussian shift model to nonparametric families of smooth densities, and from the quadratic risk to more general divergence type distances. The development relies on a connection with local proper scoring rules.

Reçu le :
Accepté le :
Publié le :
DOI : 10.1016/j.crma.2011.04.015
Ehm, Werner 1

1 Institute for Frontier Areas of Psychology and Mental Health, Wilhelmstr. 3a, 79098 Freiburg, Germany 
@article{CRMATH_2011__349_11-12_699_0,
     author = {Ehm, Werner},
     title = {Unbiased risk estimation and scoring rules},
     journal = {Comptes Rendus. Math\'ematique},
     pages = {699--702},
     publisher = {Elsevier},
     volume = {349},
     number = {11-12},
     year = {2011},
     doi = {10.1016/j.crma.2011.04.015},
     language = {en},
     url = {http://www.numdam.org/articles/10.1016/j.crma.2011.04.015/}
}
TY  - JOUR
AU  - Ehm, Werner
TI  - Unbiased risk estimation and scoring rules
JO  - Comptes Rendus. Mathématique
PY  - 2011
SP  - 699
EP  - 702
VL  - 349
IS  - 11-12
PB  - Elsevier
UR  - http://www.numdam.org/articles/10.1016/j.crma.2011.04.015/
DO  - 10.1016/j.crma.2011.04.015
LA  - en
ID  - CRMATH_2011__349_11-12_699_0
ER  - 
%0 Journal Article
%A Ehm, Werner
%T Unbiased risk estimation and scoring rules
%J Comptes Rendus. Mathématique
%D 2011
%P 699-702
%V 349
%N 11-12
%I Elsevier
%U http://www.numdam.org/articles/10.1016/j.crma.2011.04.015/
%R 10.1016/j.crma.2011.04.015
%G en
%F CRMATH_2011__349_11-12_699_0
Ehm, Werner. Unbiased risk estimation and scoring rules. Comptes Rendus. Mathématique, Tome 349 (2011) no. 11-12, pp. 699-702. doi : 10.1016/j.crma.2011.04.015. http://www.numdam.org/articles/10.1016/j.crma.2011.04.015/

[1] A.P. Dawid, S.L. Lauritzen, The geometry of decision theory, in: Proc. 2nd Int. Symp. Inf. Geom. Appl., Univ. Tokyo, 2005, pp. 22–28.

[2] W. Ehm, T. Gneiting, Local proper scoring rules of order two, Preprint, , 2011. | arXiv

[3] Gneiting, T.; Raftery, A.E. Strictly proper scoring rules, prediction, and estimation, J. Amer. Statist. Assoc., Volume 102 (2007), pp. 359-378

[4] Hendrickson, A.D.; Buehler, R.J. Proper scores for probability forecasters, Ann. Math. Statist., Volume 42 (1971), pp. 1916-1921

[5] Hyvärinen, A. Estimation of non-normalized statistical models using score matching, J. Mach. Learn. Res., Volume 6 (2005), pp. 695-709

[6] Hyvärinen, A. Optimal approximation of signal priors, Neural Computation, Volume 20 (2008), pp. 3087-3110

[7] M. Parry, A.P. Dawid, S. Lauritzen, Proper local scoring rules, Preprint, , 2011. | arXiv

[8] Stein, C.M. Estimation of the mean of a multivariate normal distribution, Ann. Statist., Volume 9 (1981), pp. 1135-1151

Cité par Sources :