This directory contains complementary files for the StanCon 2020 talk "Approximate Bayesian inference for latent Gaussian models". New research is demonstrated in the 2022 notebook. It is recommended that readers consult the more recent notebook.
This notebook is available in html format and can be viewed, without downloading the repo, via this link.
The version of cmdstan which supports the integrated Laplace approximation can be installed by following the instructions on this github repository
https://github.com/SteveBronder/laplace_testingThe notebook is available in pdf and html format: lgm_stan.pdf or lgm_stan.html.
The top of the notebook lists the R packages required to run the code.
In addition, readers can install the right version of CmdStan
using script/intall.sh.
The code to run the examples can be found in the script directory.
Contact: charles.margossian@columbia.edu