IWSM 2021: Scientific Course
"Approximate Bayesian inference for spatio-temporal models" by Virgilio Gómez-Rubio (Universidad de Castilla la Mancha, Spain) on Wednesday July 21, 2021.
- Limited to number of participants.
- Time schedule: TBA.
-
Course abstract
-
Spatio-temporal models are becoming ubiquitous due to the wealth of geo-referenced data. Bayesian inference provides a unique framework to formulate and fit these models, which can be expressed as Bayesian hierarchical models to account for the different levels of uncertainty. As these models can be highly parameterized, specific computational methods may be required for model fitting. The integrated nested Laplace approximation (INLA, Rue et al., 2009) relies on numerical approximations to estimate the posterior marginals of the latent effects and hyperparameters of hierarchical models that can be expressed as a latent Gaussian Markov random field (GRMF). INLA is particularly interesting for spatio-temporal models and several authors have already described how to use INLA to fit different types of spatio-temporal models (Krainski et al, 2019; Gómez-Rubio, 2020). This course will provide an outline of INLA for spatio-temporal modelling. Examples will be illustrated using the R programming language and the R-INLA package.
References:
- V. Gómez-Rubio (2020). Bayesian Inference with INLA. CRC Press.
- E.T. Krainski et al. (2019). Advanced Spatial Modeling with Stochastic Partial Differential Equations Using R and INLA. CRC Press.
- H. Rue, S. Martino and N. Chopin (2009). Approximate Bayesian inference for latent Gaussian models by using integrated nested Laplace approximations. JRSS-B 71(2), 319-392.