checkIsConverged: checkConvergence - Check the convergence of a mcmc.list...

In tmcd82070/EoAR: Evidence of Absence Regression (EoAR)

Description

Run various diagnostic procedures in the coda package to check convergence of a set of mcmc chains.

Usage

checkIsConverged(obj, criterion = 1.1, quiet = FALSE)

Arguments

obj	An mcmc.list object from the rjags package. These are output by the coda.samples function. This mcmc.list must contain 2 or more chains.
quiet	Whether to print any output on screen.

Details

This routine runs the Gelman and Rubin (1992) diagnostic for a set of MCMC chains by calling gelman.diag in the coda package.

Value

A list with the following components

• converged: If the uppper limits of the 'potential scale reduction factors' associated with all parameters are less than criterion, the chain is declared converged and this is TRUE. Otherwise, this component of the returns is FALSE.

• Rhats: the 'potential scale reduction factors' or Rhats associated with all parameters.

Author(s)

Trent McDonald

References

Gelman, A and Rubin, DB (1992) Inference from iterative simulation using multiple sequences, Statistical Science, 7, 457-511.

See Also

gelman.diag, checkIsConverged

Examples

## Not run: 
jags = jags.model(file="someJAGSFile.txt",
   data=someJAGS.data, inits=someJAGS.inits,
   n.chains=3,n.adapt=100)
update(jags, n.iter=1000)
out = coda.samples(jags,
   variable.names=c("someParameter"),
   n.iter=1000,
   thin=2)

conf <- chechIsConverged(out)

## End(Not run)

tmcd82070/EoAR documentation built on July 13, 2021, 5:52 p.m.