The topics organised according to content

Here I categorise the tutorial pages according to type of content. Organising the topics in a useful way is a long term goal. Please see All topics for a chronological list of everything.

Stream 1: Basic INLA

• Installation and general troubleshooting
• Application: Seeds example
• Simulation-inference Time series and non-linear effects

Stream 2: Understanding INLA

• Basic computations with precision matrix: AR1 example
• How INLA deals with hyper-parameters: 1D Reconstruction example

Stream 3: Basic SPDE

• Application: Measuring calcium content in soil samples
• Simulation-inference: A simple spatial model
• How to make a bad mesh

Stream 4: Barrier model

• Application: Barrier Model applied to the Archipelago Data
• Simulation-inference: Fish in an archipelago
• Polygons and mesh from coastlines
• Surface reconstruction with the Barrier model

Stream 5: Priors and hyper-priors

• Detailed consideration of priors in time-series: Analysing unemployment data with the AR1 Model
• Future topic: Visualising spatial priors
• Future topic: Visualising priors vs visualising hyper-priors
• Future topic: Computing results not in the prior model structure

Stream 6: Predictions and model comparison

• Multivariate predictions on held-out data - binomial example
• Requested: Predictions with Gaussian likelihood (no topic)

Stream 7: Model misspecification

• Missing covariate in Gaussian: Unknown covariates are included in the Gaussian noise
• Missing covariate in Poisson: What I think about overdispersed Poisson

Stream 8: Developing a model

This is a planned future stream. The applied models found in these, and other, tutorials are “developed products”. Any model you see was not the first model the author tried, but he/she tried several models before presenting any result. In this stream I want to have a step by step description of how you can go about the process of developing a model for a particular application.

Other material

• A set of exercises