This repository houses the R package spmrf, which is used for fitting Bayesian nonparametric adaptive smoothing models as described in Faulkner and Minin (2017). The spmrf package interfaces with Stan, which is a C++ package for performing Bayesian inference using Hamiltonian Monte Carlo (see Stan can be interfaced with the R package rstan, and thus the spmrf package depends on the rstan package to fit models.


  1. Install package dependency rstan and install package devtools using install.packages function. Note that if you do not already have rstan installed, you may need to install additional packages such as Rtools if using a Windows platform, or Xcode if you are using a Mac. See the rstan prerequisites for more information. If you want the vignettes, you may also need to install the rmarkdown package
  2. Load devtools using library(devtools).
  3. Install spmrf from GitHub using either
  4. install_github("jrfaulkner/spmrf") or
  5. install_github("jrfaulkner/spmrf", build_vignettes=TRUE) if you want the vignette documentation which provides examples of using spmrf. Note that building vignettes will make the load take a little longer.


The following vignettes provide some examples using the spmrf package with step-by-step instructions and R code.

  1. Introduction to spmrf
  2. Coal Mining Example
  3. Unequal Grid and Covariates Example


Faulkner, J. R., and V. N. Minin. 2017. Locally adaptive smoothing with Markov random fields and shrinkage priors, Bayesian Analysis, advance publication.

faulknerjam/bnps documentation built on Dec. 4, 2017, 5:28 p.m.