Small Error in documentation

[siedan17]

Hi!

I looked through the source code of the reducedorder model.
In there I found the description of the crossvalidation function. It said that if n_splits = num_of_snapshots, then it is the same as loo_error() except that here the relative error will be computed instead of the absolute (loo_error).

From my inspection, I think that both functions compute the relative error. Am I wrong?

Kind regards,
Daniel

[ndem0]

Yes, you are right, it derives from one of the previous versions!
Thanks for spotting it! I'll correct asap