For regressions, how do you know when a transformation might be useful and when it wouldn't be?

Posted Date: