(LearnerND) scale y-values

learners should be invariant under scaling, but that is not the case for the y-value (the output of the evaluated function) for the LearnerND

added LearnerND label

I think this should be considered a bug instead of an enhancement since now using a LearnerND in a BalancingLearner will not work well.

added Bug and removed Enhancement labels

mentioned in merge request !128 (merged)

closed via merge request !128 (merged)

mentioned in commit a90d2579

mentioned in issue #125 (closed)