Jeffreys priors for mixture estimation - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2017

Jeffreys priors for mixture estimation

Résumé

While Jeffreys priors usually are well-defined for the parameters of mixtures of distributions, they are not available in closed form. Furthermore, they often are improper priors. Hence, they have never been used to draw inference on the mixture parameters. We study in this paper the implementation and the properties of Jeffreys priors in several mixture settings, show that the associated posterior distributions most often are improper, and then propose a non-informative alternative for the analysis of mixtures.

Dates et versions

hal-01517565 , version 1 (03-05-2017)

Identifiants

Citer

Clara Grazian, Christian Robert. Jeffreys priors for mixture estimation. 2017. ⟨hal-01517565⟩
88 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More