Agenda

Marina Vannucci
Rice University

Abstract

In this talk I will present novel Bayesian methodologies that have been developed for specific applications to science. I will first motivate the talk by introducing data from different scientific studies, including neuroscience and microbiome studies. I will highlight common goals of these studies, such as variable selection, estimation of dependence structures and of latent states underlying the observed data. I will then describe statistical methodologies specifically tailored to these scientific problems, focusing the discussion on aspects of covariate-dependent modeling, variable selection priors, graph estimation, and strategies for efficient posterior inference. Applications will include assessing seizure risk in patients with epilepsy, estimating dynamic brain connectivity states based on neuroimaging data, and integrative modeling of microbiome data.

- - -

This lecture is included in the programme of the 2024 ISBA World Meeting

For further information please contact isba2024@unive.it