Probabilistic programming
Content
- Computational tricks (blocking, reparametrization, marginalization, etc.)
Learning objectives
At the end of the chapter, students should be able to
- diagnose performance of MCMC algorithms and implement potential remedies
Readings
- Gelman et al. (2008)
Complementary readings
Slides
View all slides in new window Download PDF of all slides
Code
- R script with Gibbs sampler and Nimble for one-way ANOVA, vs Stan code
- R script for the Tokyo rainfall example
- R script for the eight school example
- Stochastic volatility model:
- Stan code 1 and 2 and
cmdstanrR script - Nimble code
- Stan code 1 and 2 and
- Smartwatch example and
cmdstanrR script
References
Gelman, A., Dyk, D. A. van, Huang, Z., & Boscardin, J. W. (2008). Using redundant parameterizations to fit hierarchical models. Journal of Computational and Graphical Statistics, 17(1), 95–122. https://doi.org/10.1198/106186008x287337