Title: Application-driven model development in space and space-time using the INLA-SPDE approach
Abstract: In this talk I give an overview of my research before I came to UiO. First, I will guide you through some applications that have motivated our methodology. Then I introduce you to two powerful frameworks, INLA for fast Bayesian inference, and the SPDE approach for good mathematical model building. I will show how we used these frameworks to work on applications and methodology. Applications include ecology, landslides, and disease mapping. Methodology focuses on the Barrier models, and a new class of non-separable space-time models extending the spatial Matern fields. INLA: Integrated Nested Laplace Approximations. SPDE: Stochastic Partial Differential Equations.
Download the flyer here.