A0274
Title: Model selection under sample selection and model misspecification
Authors: Francisco Javier Rubio - University College London (United Kingdom) [presenting]
Adam Iqbal - Durham University (United Kingdom)
Emmanuel Ogundimu - University of Durham (United Kingdom)
Abstract: The effects of model misspecification and sample selection in Bayesian variable selection are discussed, focusing on the formulation of local and non-local priors for both the outcome and selection processes. Strategies for calibrating these priors will also be discussed. Additionally, we address computational challenges related to the non-concavity of the log-likelihood function. Theoretical and empirical results will be presented to illustrate the impact of model misspecification, along with a brief real-data example.
