diagnostics {nclSLR}R Documentation

Graphical diagnostics for arrays of MCMC output.

Description

This function generates graphical diagnostics for an array of MCMC output such as that returned by applying the function to.array to a stanfit object. For every parameter, a row of three plots is generated: a trace plot, an ACF plot and a kernel density plot. If there is output from more than one chain, the default behaviour is to overlay the plots for different chains in different colours.

Usage

diagnostics(mcmc, rows = 3, lag.max = 50, pool = FALSE,
  colours = NULL)

Arguments

mcmc

A matrix with dimensions: iterations, parameters; or a three dimensional array with dimensions: iterations, chains, parameters. The final (i.e. parameter) component of the dimnames attribute of the matrix or array should contain the parameter names.

rows

A number indicating the number of parameters to plot per page on the graphics device.

lag.max

A number indicating the maximum lag for the ACF plots.

pool

A logical. If TRUE the samples are pooled across chains before generating the plots.

colours

A vector indicating the colours to use to represent each chain. Colours can be specified using any of the three kinds of R colour specifications, i.e. a colour name (as listed by colors()), a hexadecimal string of the form "#rrggbb" or "#rrggbbaa" or a positive integer i meaning palette()[i].

Value

NULL

Examples

srs = array(rnorm(8000), c(1000, 2, 4)) # Example for illustration only!
dimnames(srs) = list(NULL, NULL, paste("theta[",1:4,"]",sep=""))
diagnostics(srs)


[Package nclSLR version 0.1.0 Index]