Méthodes de régression. II. Critères baysésiens (à suivre)

Auteur :
CAZES, P.

Description :
The generalized ridge estimator is described under assumptions of normality, as a Bayesian estimator. The properties of this estimator and its derivatives (ridge estimator, shrunken estimator, Marquardt's estimator) which are biased estimators of the regression coefficients vector, are studied together with the criterion of mean square error.


Type de document :
Article de périodique

Source :
Cahiers de l'analyse des données Paris, 1978, vol. 3, n°. 3, p. 257-268

Date :
1978