[R] Variable importance in Random Forests

Nagu thogiti at gmail.com
Thu Mar 6 00:30:59 CET 2008


Hi,

How do I get what splits were used for the top important variables.
For example, if my top most important variable is a continuous
variable, e.g. previous 12 month sales, and its range is between $0 to
$1M, how do I find what what were the split values for this variable
that were used in the random forests algorithm? How do I achieve the
same thing for a factor variable?

Thank you,
Nagu



More information about the R-help mailing list