"Wrong class for AUC"

noah977

Hi,

I'm have a very basic logistic regression process.  I want to use AUC (Area Under Curve) as my performance measurement.  RM seems to always pick the false class as "positive" and build the AUC graph that way.  This appears to be backward.  I want the true class to be positive for the AUC curve.

How can I tell RM to switch them??

Thanks,

-N