traceworstRhat {jagshelper}

Trace plots corresponding to the worst values of Rhat

Description

Trace plots with kernel densities will be created for parameters with the largest (worst) associated values of Rhat. This function is primarily intended for parameters with a vector (or array) of values.

Rhat (Gelman-Rubin Convergence Diagnostic, or Potential Scale Reduction Factor) is calculated within 'JAGS', and is commonly used as a measure of convergence for a given parameter node. Values close to 1 are seen as evidence of adequate convergence. n.eff is also calculated within 'JAGS', and may be interpreted as a crude measure of effective sample size for a given parameter node.

Usage

traceworstRhat(x, p = NULL, n.eff = FALSE, margin = NULL, parmfrow = NULL, ...)

Arguments

x

Output object returned from jagsUI

p

Optional vector of parameters to subset

n.eff

Whether to plot parameters with the smallest associated values of n.eff instead. Defaults to FALSE.

margin

In the case of a 2+ dimensional array associated with a given parameter, this will have the effect of plotting the worst Rhat corresponding to each margin specified. For example, specifying margin=2 (column) will plot the parameter with the worst Rhat value from each column. In contrast, specifying margin=NULL (the default) will cause the function to plot the single array element with the largest Rhat value.

parmfrow

Optional call to par(mfrow) for the number of rows & columns of plot window. Returns the graphics device to previous state afterward.

...

additional plotting arguments or arguments to tracedens_jags()

Value

NULL

Author(s)

Matt Tyers

References

Gelman, A., & Rubin, D. B. (1992). Inference from Iterative Simulation Using Multiple Sequences. Statistical Science, 7(4), 457–472. http://www.jstor.org/stable/2246093

See Also

plotRhats, check_Rhat, qq_postpred, ts_postpred, plot_postpred

Examples

## plotting everything
traceworstRhat(SS_out, parmfrow=c(3,2))
SS_out$Rhat  # the associated values

traceworstRhat(SS_out, parmfrow=c(3,2), n.eff=TRUE)
SS_out$n.eff  # the associated values

## in the case of a 2-D array, setting margin=2 gives the max Rhat
## associated with each column, rather than the global max
traceworstRhat(x=SS_out, p="cycle_s", margin=2, parmfrow=c(2,2))
SS_out$Rhat
traceworstRhat(x=SS_out, p="cycle_s", margin=2, parmfrow=c(2,2), n.eff=TRUE)
SS_out$n.eff