Adele Cutler and Leo Breiman
A Program for Visualizing Random Forest Classifiers

Random forest classifiers are about as accurate as support vector
machines, but have the potential to be much more interpretable. We describe
random forest classifiers and introduce visualization software that is
being developed to help with interpretation. The software is based on
the java component library VisAD (http://www.ssec.wisc.edu/~billh/visad.html).