Zero-inflated GAMs for the analysis of spatial and spatial-temporal correlated data using R-INLA

Onsite course :

Zero-inflated GAMs for the analysis of spatial and spatial-temporal correlated data using R-INLA 

 

We start with an introduction to Bayesian statistics and show how to execute Poisson and negative binomial GLMs in R-INLA. We then discuss zero-inflated GLMs for count data and continuous data, and show how to execute such models in R-INLA. In the second part of the course, we discuss generalised additive models (GAM) and show how to execute these models in R-INLA. In the third part we extend the zero-inflated GAMs with spatial, and spatial-temporal dependency. During the course, several case studies are presented, integrating statistical theory with applied analyses in a clear and understandable manner. Throughout the course, we will use the R-INLA package in R. This is a non-technical course.

Pre-required knowledge
Participants should be familiar with data exploration, linear regression and basic GLMs (i.e. Poisson and negative binomial GLM) in R. The course does contain revision/preparation material with on-demand videos.

Course content

  • Preparation material (with on-demand video):
  • Linear regression exercise in R.
  • Poisson/negative binomial GLM exercise in R.
  • Matrix notation.
  • DHARMa for model validation.

Module 1:

  • General Introduction.
  • Brief introduction to Bayesian Analysis. Conjugate priors. Diffuse versus informative priors.
  • Theory presentation on R-INLA.
  • Exercise on executing a Poisson/NB GLM in R-INLA.
  • Theory presentation on zero-inflated GLM for count data.
  • Exercise on executing a zero-inflated Poisson/NB GLM in R-INLA.

Module 2:

  • Catching up.
  • Theory presentation on hurdle models for count data and continuous data.
  • Exercise showing how to execute a zero-altered Poisson (or NB) GLM for the analysis of zero-inflated count data.
  • Exercise comparing Tweedie and zero-altered Gamma GLM for the analysis of zero-inflated continuous data.

Module 3:

  • Theory presentation on GAM.
  • Exercise on executing (zero-inflated) Poisson and negative binomial GAMs in R-INLA.
  • Theory presentation on adding spatial correlation to a regression model in R-INLA.

Module 4:

  • Catching up.
  • Exercise on adding spatial correlation to a zero-inflated Poisson or negative binomial GAM.
  • Exercise on adding spatial correlation to a Tweedie GAM for the analysis of zero-inflated continuous data.
  • Theory presentation on adding spatial-temporal correlation to a GLM in R-INLA.

Module 5:

  • Exercise on adding spatial-temporal correlation to a Poisson or negative binomial GAM.
  • Exercise on adding spatial-temporal correlation to a Poisson or negative binomial GAM.
  • Time allowing: More exercises.

We reserve the right to change the exercises. Pdf files of all theory material will be provided. All exercises consist of data sets and annotated R scripts. Access to the course website is for 12 months. The Module 1 - Module 5 material does not contain on-demand video.

For terms and conditions, see:
https://www.highstat.com/index.php/component/hikashop/checkout/termsandconditions/step-3/pos-6/tmpl-component