Diagnostic plots for HMC and NUTS using ggplot2.

stan_diag(object,
            information = c("sample","stepsize", "treedepth","divergence"),
            chain = 0, ...)
  stan_par(object, par, chain = 0, ...)

  stan_rhat(object, pars, ...)
  stan_ess(object, pars, ...)
  stan_mcse(object, pars, ...)

Arguments

object

A stanfit or stanreg object.

information

The information to be contained in the diagnostic plot.

par,pars

The name of a single scalar parameter (par) or one or more parameter names (pars).

chain

If chain=0 (the default) all chains are combined. Otherwise the plot for chain is overlaid on the plot for all chains combined.

...

For stan_diag and stan_par, optional arguments to arrangeGrob. For stan_rhat, stan_ess, and stan_mcse, optional arguments to stat_bin in the ggplot2 package.

Value

For stan_diag and stan_par, a list containing the ggplot objects for each of the displayed plots. For stan_rhat, stan_ess, and stan_mcse, a single ggplot object.

Details

stan_rhat,stan_ess,stan_mcse

Respectively, these plots show the distribution of the Rhat statistic, the ratio of effective sample size to total sample size, and the ratio of Monte Carlo standard error to posterior standard deviation for the estimated parameters. These plots are not intended to identify individual parameters, but rather to allow for quickly identifying if the estimated values of these quantities are desireable for all parameters.

stan_par

Calling stan_par generates three plots: (i) a scatterplot of par vs. the accumulated log-posterior (lp__), (ii) a scatterplot of par vs. the average Metropolis acceptance rate (accept_stat), and (iii) a violin plot showing the distribution of par at each of the sampled step sizes (one per chain). For the scatterplots, red points are superimposed to indicate which (if any) iterations encountered a divergent transition. Yellow points indicate a transition that hit the maximum treedepth rather than terminated its evolution normally.

stan_diag

The information argument is used to specify which plots stan_diag should generate:

  • information='sample'Histograms of lp__ and accept_stat, as well as a scatterplot showing their joint distribution.

  • information='stepsize'Violin plots showing the distributions of lp__ and accept_stat at each of the sampled step sizes (one per chain).

  • information='treedepth'Histogram of treedepth and violin plots showing the distributions of lp__ and accept_stat for each value of treedepth.

  • information='divergence'Violin plots showing the distributions of lp__ and accept_stat for iterations that encountered divergent transitions (divergent=1) and those that did not (divergent=0).

Note

For details about the individual diagnostics and sampler parameters and their interpretations see the Stan Modeling Language User's Guide and Reference Manual at http://mc-stan.org/documentation/.

See also

Examples

if (FALSE) { fit <- stan_demo("eight_schools") stan_diag(fit, info = 'sample') # shows three plots together samp_info <- stan_diag(fit, info = 'sample') # saves the three plots in a list samp_info[[3]] # access just the third plot stan_diag(fit, info = 'sample', chain = 1) # overlay chain 1 stan_par(fit, par = "mu") }