Outlier in Cross Validation

How can we justify reason about outlier in cross validation for example I have test results [0.98,0.95,0.98,0.64]

If for example, these are from a classification model, in the outlier fold, are there samples with a particular class in validation that are not in train? (or it might be that fold is extremely unbalanced class-wise). But in the first instance I would be looking for a missing class or two in that train fold

Thanks @AllenK and sorry for late response

1 Like