This is the course materials page and package for the Edinburgh inlabru course, 19 and 20 May 2025. Follow the menu links to information and tutorials, as and when they become available.
Event Overview
Bayesian Latent Gaussian Models (LGMs) are closely related to Generalized Additive Models (GAMs), offering Bayesian estimation and uncertainty quantification for spatial and spatio-temporal models. The INLA and inlabru R packages combine these Gaussian process models with numerical optimization and integration techniques, in a fast and flexible analysis toolkit. The taught part of the course will provide an overview of LGM theory and the INLA/inlabru methods and software, while the hands-on sessions will make sure the attendees will be ready to start doing spatial LGM modelling in R as soon as the course is over.
Topics covered
Basics of latent Gaussian process models in the Bayesian spatial statistics context. The principles of the INLA method for fast Baysian inference, and inlabru extensions for non-linear models. The inlabu package principles and interface. Building spatial and spatio-temporal models for point-referenced, spatially aggregated, and point pattern observations. Computing and assessing posterior predictions and visualisation. Diagnosing modelling problems.
Instructor
Prof Finn Lindgren is Chair of Statistics in the School of Mathematics, University of Edinburgh.
His research focuses on spatial and spatio-temporal stochastic models, environmetrics, and computational methods and software. Among many others, he co-authored the influential paper “An Explicit Link Between Gaussian Fields and Gaussian Markov Random Fields: The Stochastic Partial Differential Equation Approach,” published in the Journal of the Royal Statistical Society: Series B. Professor Lindgren has contributed to the development of several R packages, including INLA for Bayesian latent Gaussian models and inlabru, a user-friendly interface for INLA with additional features
Learning outcomes
- Understand the basic theory underpinning spatial latent Gaussian process models and Bayesian inference
- Use inlabru to fit various spatial models to data, including point-referenced, aggregated, and point pattern data.
- Be able to compute posterior predictions.
- Assess and compare models
Daily timetable (tentative)
Monday 19/5
- 09:30 - 10:30 Lecture (Spatial modelling with random fields)
- 10:30 - 11:00 coffee break
- 11:00 - 12:30 Lecture/Hands on session (Introduction to
INLA/inlabru/fmesher) - 12:30 - 13:30 Lunch
- 13:30 - 15:00 Lecture/Hands on session (Spatial models for point-referenced data)
- 15:00 - 15:30 break
- 15:30 - 17:00 Lecture/Hands on session (Aggregated counts and non-linear predictors) (Non-separable space-time)
Tuesday 20/5
- 09:30 - 10:30 Lecture (Point process models) (Poisson Point processes) (Distance sampling transect surveys)
- 10:30 - 11:00 coffee break
- 11:00 - 12:30 Lecture/Hands on session (Spatial covariates) (Space-time)
- 12:30 - 13:30 Lunch
- 13:30 - 15:00 Lecture/Hands on session (Multi-likelihood models; hurdle model example)
- 15:00 - 15:30 break
- 15:30 - 17:00 Lecture/Hands on session (Predictive model assessment) (Spatially varying coefficients)