Open tballison opened 5 years ago
If each feature could report a complexity figure, we could roll those up to the experiment level, and users would be able to select the "simplest" experiment that is not significantly different from the best performer.
If each feature could report a complexity figure, we could roll those up to the experiment level, and users would be able to select the "simplest" experiment that is not significantly different from the best performer.