| pairs.mcmc.list | R Documentation |
A matrix of scatterplots is produced.
## S3 method for class 'mcmc.list'
pairs(x, n = 25, col = 1:length(x),
col.hist = "gold", col.image = terrain.colors(50),
density = TRUE, contour = TRUE, mean = TRUE, ...)
x |
an 'mcmc.list' object. |
n |
number of of grid points in each direction for two-dimensional kernel density estimation. Can be scalar or a length-2 integer vector. |
col |
color for chains in upper panel scatterplots. |
col.hist |
color for histogram fill in diagonal panels. |
col.image |
color palette for image plot in lower panel scatterplots. |
density |
logical, if image plot based on the two-dimensional kernel density estimation should be plotted in lower panel. |
contour |
logical, if contour plot based on the two-dimensional kernel density estimation should be plotted in lower panel. |
mean |
logical, if lines should indicate means of the posterior densities in the panels. |
... |
additional graphical parameters/arguments. |
The function produces a scatterplot matrix for 'mcmc.list' objects.
Diagonal panels are posterior densities with labels and
rug on the top. Upper panels are pairwise bivariate scatterplots
with coloring corresponding to chains, thus highlighting mixing properties
although not as clearly as trace plots. Lower panels
are two-dimensional kernel density estimates based on
kde2d function
of MASS package using image
and contour.
The function returns NULL invisibly and
produces a plot as a side effect.
Peter Solymos, solymos@ualberta.ca
pairs, plot.mcmc.list
Two-dimensional kernel density estimation:
kde2d in MASS package
data(regmod) pairs(regmod)