Introduction to JAGS

Introduction to JAGS

From http://mcmc-jags.sourceforge.net/ “JAGS is Just Another Gibbs Sampler. It is a program for analysis of Bayesian hierarchical models using Markov Chain Monte Carlo (MCMC) simulation…””

Installation

brew install jags

Then for R you need first in ~/.R/Makevars set correct CC and CXX fo JAGS:

CC=clang
CXX=clang++

Then start R session and install rjags:

install.packages("rjags")

Usage

Load package

library('rjags')

Specify model

model.str = "model {
    for (i in 1:n) {
        y[i] ~ dnorm(mu, 1.0/sigma2)
    }
    mu ~ dt(0.0, 1.0/1.0, 1)
    sigma2 = 1.0
}"

Set up the model

y <- c(1.2, 1.4, -0.5, 0.3, 0.9, 2.3, 1.0, 0.1, 1.3, 1.9)
n <- length(y)

data.jags <- list(y=y, n=n)
params <- c("mu")

inits <- function() {
  inits <- list("mu"=0.0)
}

model <- jags.model(textConnection(model.str), data=data.jags, init=inits)

Run sampler

update(model, 500)

model.simulation  = coda.samples(model=model, variable.names=params, n.iter=1000)

Post processing

library("coda")

plot(model.simulation)

Updated: 2020-12-03