Model Selection: Some Remarks from a Bayesian Viewpoint

Florens, J-P;Mouchart, Michel
(1983) 4th Franco-Belgian Meeting of Statisticians — Location: Louvain - la - Neuve (24.November.1983)

Files

No attached file found for this publication.

Details

Authors
  • Florens, J-PUniversité d'Aix Marseille
    Author
  • Mouchart, MichelUCLouvain
    Author
Abstract
Model selection is approached in the framework of a Bayesian experiment with a discrete sub-σ-field on the parametrer space. A function generating such a sub-σ-field is called a label model. We consider the Bayesian estimability of the model label, some comparison between two experiment conditional to a model label and some comparison between two experiments, each conditional to a model label and some comparison between two experimetns, each one conditional to different model labels.The analysis is partly asymptotic.
Affiliations

Citations

  • APA
  • Chicago
  • FWB

Florens, J.-P., & Mouchart, M. (1983). Model Selection: Some Remarks from a Bayesian Viewpoint. In J-P Florens, Michel Mouchart, J-P Raoult, Léopold Simar (ed.), Model Choice : Proceedings of the 4th Franco-Belgian Meeting of Statisticians (p. p. 27-44). Publications des Faculté Universitaires Saint-Louis. https://hdl.handle.net/2078.5/65184