trace.MCmcmc {MethComp}

Functions to graphically assess the convergence of the MCMC-simulation in a MCmcmc object

Description

These functions display traces for the relevant subset of the parameters in a MCmcmc object.

Usage

## S3 method for class 'MCmcmc'
trace(
  obj,
  what = "sd",
  scales = c("same", "free"),
  layout = "col",
  aspect = "fill",
  ...
)

Arguments

obj

A MCmcmc object.

what

Character indicating what parameters to plot. Possible values are "sd" or "var" which gives plots for the variance components (on the sd. scale), "beta" or "slope", which gives plots for slope parameters and "alpha" or "int", which gives plots for the intercept parameters.

scales

Character vector of length two, with possible values "same" or "free", indicating whether x- and y-axes of the plots should be constrained to be the same across panels. For pairs only the first element is used to decide whether all panles should have the same axes.

layout

Character. If "col" parameters are displayed columnwise by method, if "row" they are displayed row-wise.

aspect

How should the panels be scaled. Default ("fill") is to make a panels take up as much place as possible.

...

Further aruments passed on to the Lattice

Details

A Lattice plot is returned, which means that it must printed when these functions are called in a batch program or inside another function or for-loop.

trace plots traces of the sampled chains, post plots posterior densities of the parameters and pairs plots a scatter-plot matrix of bivariate marginal posterior distributions.

Value

A Lattice plot.

Author(s)

Bendix Carstensen, Steno Diabetes Center, [e-mail], http://BendixCarstensen.com.

See Also

MCmcmc, plot.MCmcmc, ox.MC, sbp.MC

Examples

# Load a provided MCmcmc object
data( ox.MC )
trace.MCmcmc( ox.MC, what="beta" )
pairs( ox.MC, what="sd" )