Improved discussion on leakage

[bgreenwell]

Great resource: https://reproducible.cs.princeton.edu/

Really ties into preprocessing before data splitting and dealing with class imbalance!

[bradleyboehmke]

In the ed. 1 we talked about leakage in the feature engineering section: https://bradleyboehmke.github.io/HOML/engineering.html#data-leakage

Are you thinking keep it in that same location or move/expand to elsewhere?

[bgreenwell]

Definitely expanding, and maybe reference the idea of “model info sheets”: https://reproducible.cs.princeton.edu/. I also have a few simulations illustrating the impact of leakage in the preprocessing stage (like feature selection followed by cross validation and up sampling).

[bgreenwell]

Illegitimate features is another one we don’t really discuss, but seems to be a common issue.