Session 8
Group Discussion
Practicals
INLA
The integrated nested Laplace approximation (INLA) is a method for approximate Bayesian inference. It is a faster alternative to Markov chain Monte Carlo, available via the R-INLA package.
- website: R-INLA
Communication
Key things to report in a paper
- The data
- methods, key observational uncertainties
- The model
- \(f\) function(s), \(x\) variables, \(\theta\) parameters
- Priors
- All parameters or a subset? What joint probability distribution did you assign? What information did you use to quantify it: literature, personal judgment, expert elicitation?
- Likelihood function
- what likelihood function did you assign? How did you account for stochastic, systematic and representativeness errors?
- MCMC
- Which algorithm did you use? How many chains and iterations? How did you assess convergence?
- Posterior distribution for the parameters
- What were posterior modes, means, variances and major correlations? How different was the posterior from the prior?
- Posterior predictions
- when using the posterior probability distribution for the parameters, how well did the model(a) reproduce data used in the calibration, (b) predict data not used in the calibration?
See Chapter 11 in Oijen, M. van. (2020). Bayesian Compendium. Springer International Publishing.
Isotopes
SIBER: Stable Isotope Bayesian Ellipses in R
- website: SIBER on CRAN
- Read: Paper in J. Animal Ecology
BACON: Radiocarbon Isotope Age-Depth Modelling using Bayesian Statistics
Bayesian Software Decision Tree
Back to course schedule ⏎