Recently I read in Foreman and Scholz 2009
) that Weka
Explorer and Experimenter use different methods to calculate AUC in
cross validation, at least for version 3.6.
Is that still the case for 3.7.13? If so, could someone make a statement
on what is the rationale behind this diverging approaches, and whether
there is some possibility to force the use of a particular (i.e.