OBsMD: an R package for objective Bayesian model discrimination in follow-up design

Laura Deldossi; Marta Nai Ruscone

Conference proceeding

OBsMD: an R package for objective Bayesian model discrimination in follow-up design

Laura Deldossi and Marta Nai Ruscone

Sco 2013: proceedings of the 8th complex data modeling and computationally intensive statistical methods for estimation and prediction, Milano, Italy, pp.1-4

Sco 2013: 8th complex data modeling and computationally intensive statistical methods for estimation and prediction (Milano, 09/09/2013 - 11/09/2013)

2013

Abstract

Occasionally screening designs do not lead to unequivocal conclusions regarding which combinations of factors (models) are active. From a Bayesian viewpoint, this means that the posterior distribution on model space will be fairly evenly spread out over a few models. In these circumstances, a follow-up design is needed, the aim being of choosing extra runs in order to solve, or at least alleviate, this ambiguity. This R-package OBsMD implements the objective Bayesian methodology developed in Consonni and Deldossi (2013) for this scope.

Files and links (1)

pdf

bitstream_7bc790cd-3545-453b-a647-0fd0338876b790.55 kB

Ask the Library / Chiedi alla Biblioteca Restricted Access

Metrics

26 Record Views

Details

Title: OBsMD: an R package for objective Bayesian model discrimination in follow-up design
Creators: Laura Deldossi (Author)
Marta Nai Ruscone (Author)
Publication Details: Sco 2013: proceedings of the 8th complex data modeling and computationally intensive statistical methods for estimation and prediction, Milano, Italy, pp.1-4
Date published: 2013
Conference: Sco 2013: 8th complex data modeling and computationally intensive statistical methods for estimation and prediction (Milano, 09/09/2013 - 11/09/2013)
Identifiers: 9788864930190
Format: A stampa
Language: English
Copyright: Tutti i diritti sono riservati ai legittimi detentori del copyright. L'Università Carlo Cattaneo - LIUC pubblica i dati relativi alle pubblicazioni di ricerca realizzate dai propri affiliati. La presenza nell'archivio ARL di testi completi non determina in alcun modo la libera riproduzione degli stessi, ma esclusivamente la possibilità della loro consultazione sul sito di ARL.
Academic Unit: Area di ricerca Scienze economiche e statistiche
Resource Type: Conference proceeding
Identifiers: 991000853616705126