Thread Subject: crossvalind" function in the bioinformatics toolbox:

For the first example of "crossvalind" function in the
bioinformatics toolbox:
       % Create a 10-fold cross-validation to compute
classification error.
       load fisheriris
       indices = crossvalind('Kfold',species,10);
       cp = classperf(species);
       for i = 1:10
           test = (indices == i); train = ~test;
           class =
classify(meas(test,:),meas(train,:),species(train,:));
           classperf(cp,class,test)
       end

Is the output of cp.CorrectRate the average of the 10-fold
cross validation or only the CorrectRate of the last fold?
If it is the avearage of the ten folds, I suggest to add one
more property of cp, that is, "the variance of the
CorrectRate". Only Mean and variance together would make a
statistical significance