Good Afternoon,
I have not understand why the last choice of question 3 of evaluation hands-on (the one about another guarantee of goodness of kappa score) it is better than the other two (expecially wrt to the second choice).
Thanks a lot!
answer 1 is non-sense and answer 2 is a much weaker/poorer version of answer 3; pretty much the same difference as in 1xN versus KxM cross-validation (with N=KxM).
Note (FYI, out of course scope): this later (asnwer 3) relates to boostrap methods in statistics.
Note (FYI, out of course scope): this later (asnwer 3) relates to boostrap methods in statistics.
