Uses the image function to identify significant differences between
classical and robust correlation matrix estimates.
USAGE:
image.cov(object, probs=c(0.95, 0.99), ...)
REQUIRED ARGUMENTS:
object
a
fit.models object containing a
cov element and
covRob element.
OPTIONAL ARGUMENTS:
probs
a vector of probabilities determining how differences are colored in the image.
Differences with significance between 0 and min(probs) are white, differences
between max(probs) and 1 are black, and intermediate values are gray
scaled. The number of colors depends on the length of probs.
SIDE EFFECTS:
An image is produced on a new graphsheet (note that the graphsheet device
is available only on the Windows platform).