Depends on the program, of course. You're asking how robust the trained model is to outliers and how effective it is at extrapolation. Different training algori

Depends on the program, of course. You're asking how robust the trained model is to outliers and how effective it is at extrapolation. Different training algorithms have different behaviors.