Plot MCMC marginals

Plot MCMC marginals

marginalPlot(mat, thin = "auto", scale = NULL, best = NULL,
  histogram = TRUE, plotPrior = TRUE, priorTop = FALSE,
  nDrawsPrior = 1000, breaks = 15, res = 500, singlePanel = FALSE,
  dens = TRUE, col = c("#FF5000D0", "#4682B4A0"), lwd = par("lwd"), ...)

Arguments

mat

object of class "bayesianOutput" or a matrix or data frame of variables
thin

thinning of the matrix to make things faster. Default is to thin to 5000
scale

should the results be scaled. Value can be either NULL (no scaling), T, or a matrix with upper / lower bounds as columns. If set to T, attempts to retrieve the scaling from the input object mat (requires that this is of class BayesianOutput)
best

if provided, will draw points at the given values (to display true / default parameter values). Value can be either NULL (no drawing), a vector with values, or T, in which case the function will attempt to retrieve the values from a BayesianOutput
histogram

Logical, determining whether a violin plot or a histogram should be plotted
plotPrior

Logical, determining whether the prior should be plotted in addition to the posteriors. Only applicable if mat is an object of class "bayesianOutput"
priorTop

Logical, determining whether the prior should be plotted top (TRUE) or bottom (FALSE)
nDrawsPrior

Integer, number of draws from the prior, when plotPrior is active
breaks

Integer, number of histogram breaks if histogram is set to TRUE
res

resolution parameter for violinPlot, determining how many descrete points should be used for the density kernel.
singlePanel

Logical, determining whether all histograms/violins should be plotted in a single plot panel or in separate panels.
dens

Logical, determining wheter an density overlay should be plotted when 'histogram' is TRUE
col

vector of colors for posterior and prior
lwd

line width of the violin plots
...

additional parameters to pass on to the getSample

References

tracePlot correlationPlot

Examples

dat = generateTestDensityMultiNormal(n = 100000, sample = TRUE)
marginalPlot(dat(10000))
#> Warning: Parameter 'mat' is not of class 'bayesianOutput', set plotPrior to FALSE.