Mark:
> I realize one can use use a training map to cluster like features,
> but is there a way to have a "leftover" class that throws everything
> else that doesnt match a defined class into this "leftover" category?
ie you want an "outliers" class?
Hamish