plot_MCMC: Plot MCMC trajectories and posterior distributions

View source: R/plot_MCMC.R

plot_MCMCR Documentation

Plot MCMC trajectories and posterior distributions


This function uses the output of rjags::jags.model to visualise the traces of the MCMC and the corresponding densities. In particular it displays the posterior distributions of the age, if it is calculated, palaeodose and the equivalent dose dispersion parameters of the sample. The function output is very similar to plot output produced with the 'coda' package, but tailored to meet the needs in the context of the 'BayLum' package.


  sample_names = NULL,
  variables = c("A", "D", "sD"),
  axes_labels = c(A = "Age (ka)", D = "D (Gy)", sD = "sD (Gy)"),
  n.chains = NULL,
  n.iter = 1000,
  smooth = FALSE,
  rug = TRUE,
  plot_single = FALSE,



coda::mcmc.list or coda::mcmc (required): Output generated by rjags::jags.model, e.g., in Age_Computation


character (optional): Names of the used samples. This argument overrides the optional argument mtext.


character (with default): Variables in your coda::mcmc object to be plotted.


character (with default): Axes labels used for the trace and density plots. The labels should be provided as named character vector with the parameter names as the names used to assign the axes labelling. The labelling for the x-axis (trace plots) and y-axis (density plot) cannot be modified.


numeric (optional): Set the number of chains to visualise, if nothing is provided the number of chains is determined from the input object


integer (with default): Set the number of iterations to be visualised in the trace plots, regardless of the size of the input dataset as long as the real number of iterations is > n.iter. Please note that large numbers impact the plot performance.


logical (with default): Enable/disables smooth of trace plots using stats::smooth


logical (with default): Enable/disables rug under density plots


logical (with default): Enables/disables the single plot mode of the function, i.e. if set to TRUE every plot is returned in a single plot and own par settings can be applied.


further arguments that can be passed to modify the plot output. Supported arguments are lwd, lty, col, type, cex,mtext, cf. mtext for mtext and plot.default for the other arguments.


The function is used in the function Age_Computation, AgeS_Computation and Palaeodose_Computation, but can be used also as standalone plot function.


Two plots: Traces of the MCMC chains and the corresponding density plots. This plots are similar to coda::traceplot and coda::densplot.

Function version


How to cite

Kreutzer, S., Christophe, C., 2022. plot_MCMC(): Plot MCMC trajectories and posterior distributions. Function version 0.1.4. In: Christophe, C., Philippe, A., Kreutzer, S., Guerin, G., 2022. BayLum: Chronological Bayesian Models Integrating Optically Stimulated. R package version 0.2.1.


Sebastian Kreutzer, Geography & Earth Sciences, Aberystwyth University (United Kingdom). This function is a re-written version of the function 'MCMC_plot()' by Claire Christophe

See Also

Age_Computation, AgeS_Computation, Palaeodose_Computation, rjags::coda.samples and rjags packages.


data(MCMCsample,envir = environment())
object <- coda::as.mcmc(MCMCsample)

BayLum documentation built on March 18, 2022, 5:50 p.m.