A0690
Title: The Bayesian discrepancy measure: A new method for Bayesian inference
Authors: Mara Manca - University of Cagliari (Italy) [presenting]
Francesco Bertolino - University of Cagliari (Italy)
Silvia Columbu - University of Cagliari (Italy)
Monica Musio - University of Cagliari (Italy)
Abstract: The aim is to construct an index that, in the Bayesian context, allows to check the conformity of a given hypothesis with respect to the available information (prior distribution and data). The proposed evidence measure, called Bayesian Discrepancy Measure (BDM), has properties of consistency and invariance. After presenting the BDM and the related Bayesian Discrepancy Test (BDT), we show their conceptual and interpretative simplicity that allows us to easily deal with complex case studies that have not yet been addressed in the literature. Theoretical and computational developments of the BDM in more general contexts, such as model selection are also at an advanced stage.
