Introduction to GAMM and GLMM. With GAM applications to spatial, and spatial-temporal data

Onsite course :

Introduction to GAMM and GLMM with R - With GAM applications to spatial, and spatial-temporal data -

 

This course offers an introduction to generalised additive models (GAMs), linear mixed-effects models, generalised linear mixed-effects models (GLMMs), the combination of GAMs and GLMMs, and also includes examples of GAMMs applied to spatial and spatio-temporal data.

The course begins with a review of multiple linear regression, followed by a non-technical introduction to GAMs. Through a series of exercises, we demonstrate the application of GAMs to accommodate non-linear covariate effects.

The second part of the course introduces linear mixed-effects models, which are suitable for data with complex structures, such as observations from clustered or hierarchical sources (e.g., multiple measurements from the same subject). These models extend traditional models by incorporating random effects to account for the non-independence of data points.

Once familiar with mixed models, we will integrate them with GAMs, resulting in generalised additive mixed-effects models (GAMMs). We will do various exercises using GLMMs and GAMMs.

In the third part of the course, we will apply GAMs and GAMMs to various spatial and spatio-temporal datasets. Throughout the course, we will utilise GAMMs and GLMMs on different types of data (continuous and count), employing Gaussian, Poisson, and negative binomial distributions.

Course content

Module 1

  • General introduction.
  • Revision exercise on multiple linear regression.
  • Theory presentation on GAM.
  • Two introductory GAM exercises.

Module 2

  • GAM exercise discussing model selection and smoother interactions.
  • Theory presentation on linear-mixed effects models.
  • One exercise using linear mixed-effects models.
  • One exercise on Gaussian additive mixed-effects models (GAM with random effects).

Module 3

  • Time allowing: Exercise using hierarchical GAMs (the GAM equivalent of random slope models).
  • Two revision exercise on Poisson and negative binomial GLMs.
  • One exercise on negative binomial GAM.

Module 4

  • One exercise on Poisson GLMM.
  • Time allowing: One exercise on negative binomial GLMM.
  • Two exercises on Poisson and negative binomial GAMM.

Module 5

  • One exercise on (Gaussian) GAM applied to spatial data.
  • One exercise on GAM applied to spatial count data.
  • One exercise on GAM(M) applied to spatial-temporal count data.
  • Time allowing: Guidance for the analysis of binary, continuous, proportional and continuous data (Bernoulli, binomial, beta, Tweedie and Gamma distributions).
  • We will predominately use the R packages mgcv and glmmTMB for the exercises.

PRE-REQUIRED KNOWLEDGE:
Working knowledge of R, data exploration, linear regression and GLM (Poisson, negative binomial). This is a non-technical course.

The course website provides preparatory materials, including on-demand videos and R scripts covering multiple linear regression, basic matrix notation, generalised linear models, model validation using DHARMa, and the explanation of variograms. If you are not familiar with these methods, please review them before the course begins.

FREE 1-HOUR FACE-TO-FACE MEETING
The course fee includes a 1-hour face-to-face meeting with one or both instructors. You can discuss your own data, but we strongly advice that the statistical topics are within the content of the course. The 1-hour consultancy needs to be consumed in one sessions, and will take place at a mutual convenient time. It is not transferable. The meetings needs to take place within 12 months after the last live zoom module.