Sequential Monte Carlo for latent Gaussian Markov random fields
This R package contains supplementary material for paper 'Graphical model inference: Sequential Monte Carlo meets deterministic approximations' by Lindsten, Helske, Vihola (NIPS 2018).
The package contains functions for finding Gaussian approximation for Binomial CAR model, bootstrap particle filter and twisted SMC algorithms for latent field and likelihood estimation as well as delayed acceptance MCMC based on Gaussian approximation + SMC.
Installation
You can install the package from Github as follows:
devtools::install_github("helske/particlefield")