The V4.0 version of random forests contains some modifications and
major additions to Version 3.3. The additions are :
a) replacement of missing values.
b) a method to balance error in class unbalanced data sets.
c) data that can be used to see how the variables relate to the classification.
d) efficient handling of categoricals with a large number of values.
The basics of this program works are in the paper "Random Forests"
Its available on the same web page as this manual. It was recently
published in the Machine Learning. Journal. Please report bugs
either to Leo Breiman (leo@stat.berkeley.edu) or Adele Cutler
(adele@sunfs.math.usu.edu)