The functions provide visualizations of MCMC results obtained for
drift_dm() objects. Users won't call the functions directly.
Instead, they are called via plot.mcmc_dm(), and the
following documentation helps to define optional arguments that can be
passed on via ....
plot_mcmc_trace(
chains,
col_palette = grDevices::rainbow,
col_chains = NULL,
which_prms = NULL,
xlab = NULL,
ylab = NULL,
xlim = NULL,
ylim = NULL
)plot_mcmc_marginal(
chains,
col_line = NULL,
col_shade = NULL,
which_prms = NULL,
xlab = NULL,
ylab = NULL,
xlim = NULL,
ylim = NULL
)
plot_mcmc_auto(
chains,
lags = 1:30,
col_line = NULL,
which_prms = NULL,
xlab = NULL,
ylab = NULL,
xlim = NULL,
ylim = NULL,
type = NULL,
main = NULL
)
These functions are called for their side effects (producing plots).
They return NULL invisibly.
an array of MCMC samples with three dimensions
(parameters × chains × iterations; see also plot.mcmc_dm()).
This argument is not optional and will be provided by
plot.mcmc_dm().
a function to generate a color palette for chains
(default is grDevices::rainbow). Must be callable like this
col_palette(n).
a character vector, defining colors(s) to override
col_palette for chain lines. Can be a single value or a vector matching
the number of chains. Recycled if necessary to match the number of chains.
a regular expression (string), used to select a subset
of parameters to plot. For example, "^v" would match all parameters
starting with v. See also the examples in plot.mcmc_dm().
character vector(s), specifying axis labels for each panel. Recycled if necessary to match the number of plots being generated.
list(s), containing length-2 numeric vectors specifying x- and y-axis limits, respectively. Recycled if necessary to match the number of plots being generated.
a character vector, defining color(s) for the line outlines in the density plots. Can be a single value or a vector matching the number of parameters to plot. Recycled if necessary to match the number of parameters being plotted.
a character vector, defining color(s) for the shaded areas in the density plots. Can be a single value or a vector matching the number of parameters to plot. Recycled if necessary to match the number of parameters being plotted.
a numeric vector, giving the lags at which to calculate
autocorrelation (see also coda::autocorr.diag() and coda::autocorr()).
a character, specifying how autocorrelations shall be displayed.
Defaults to "h" (see the type argument of base::plot()). Recycled if
necessary to match the number of parameters being plotted.
character vector, specifying the labels above each autocorrelation plot. Defaults to parameter labels. Recycled if necessary to match the number of parameters being plotted.
plot_mcmc_trace() plots one panel per parameter, with lines for each
chain showing how values evolve over iterations. This is useful for
diagnosing convergence and mixing.
plot_mcmc_marginal() plots smoothed marginal posterior densities
collapsed over chains and iterations for each parameter, useful for
inspecting posterior distributions.
plot_mcmc_auto() plots the autocorrelation among samples as function
of lags.
plot.mcmc_dm()