Hello
I have tested KrippendorffAlphaUnitizingAgreement with a borderline case: two annotators have annotated all the text with one category.
In this case, Krippendoff's Alpha for unitizing is incalculable as chance disagreement is null.
Yet, DKpro computes a value (0,25), do you have any explanation?
Thank you