Rather than complaining when only one class is seen, wouldn't a more robust approach be to predict the single class seen with probability one? At least, in the MLJ interface, where we track the complete target pool, this shouldn't pose any issues.
One hits this corner case when subsampling binary data with a small training set, unbalanced data, and/or lots of folds.
Rather than complaining when only one class is seen, wouldn't a more robust approach be to predict the single class seen with probability one? At least, in the MLJ interface, where we track the complete target pool, this shouldn't pose any issues.
One hits this corner case when subsampling binary data with a small training set, unbalanced data, and/or lots of folds.
See also https://github.com/JuliaAI/CatBoost.jl/pull/20#issuecomment-1429286204.