\name{RNC}
\alias{RNC}
\alias{maxRNC}
%- Also NEED an '\alias' for EACH other topic documented here.
\title{ Brooks-Gelman-Rubin R, Effective Sample Size, and Autocorrelation}
\description{
  For a parameter set creates a list of the three most useful statistics
  for judging sufficiency of the MCMC sample:  \code{R} --- Brooks,
  Gelman Rubin multivariate potential scale reduction factor, \code{N}
  --- the effective sample size, and \code{C} --- the lag 5
  autocorrelation.  The function \code{maxRNC} provides the maximum
  values for those statistics across its arguments.
}
\usage{
RNC(ps)
maxRNC(pslist)
}
\arguments{
  \item{ps}{A parameter set (or other coda \code{mcmc}) object}
  \item{pslist}{A list of parameter set (or other coda \code{mcmc})
    objects} 
}
\details{
  The function \code{RNC} applies the functions \code{mpsrf} with
  arguments \code{transform=TRUE} and \code{autoburnin=FALSE},
  \code{effectiveSize}, and \code{autocorr.diag} with \code{lag=5} to the
  \code{ps} argument.   Returns the three diagnostics as a list.

  The function \code{maxRNC()} applies \code{RNC()} to each of its
  arguments and then returns the maximum of the \code{R} and \code{C}
  components and the minimum of the \code{N} component (where the
  maximium is done separately for each component).

  These three statistics are the generally most useful when looking at
  the convergence of the chain.  The \code{R} component provides a
  useful diagnostic for convergence, the \code{N} component provides a
  guide as to the sufficiency of the sample, and the \code{C} component
  helps identify problems with slow mixing.

}
\value{
  \item{R}{The Brooks-Gelman-Rubin multivariate potential scale
    reduction factor.}
  \item{N}{The effective sample size of the series.}
  \item{C}{The lag 5 autocorrelation of the series.}
}
\author{Russell Almond}
\seealso{\code{\link{mpsrf}}, \code{\link[coda]{effectiveSize}},
  \code{\link[coda]{autocorr.diag}} }
\examples{
StatShop <- Sys.getenv("STATSHOP")
if (StatShop == "") {
  warn("StatShop not installed, skipping test.")
} else {
asp.dbConnect("parameter_recovery_test","junit")
## Get information about the student model.
smname <- listStudentModels()[1]
pname <- listParameterNames(smname)

## Look at the first parameter in the student model.
RNC(getParameterVector(smname,pname[1]))

## Look at all prameters in the first model
maxRNC(getAllParameters(listModels()[1]))

## Look at all link models
dc <- daCapo()
maxRNC(unlist(dc$linkModels,recursive=FALSE))

}
}
\keyword{ts}% __ONLY ONE__ keyword per line