Fig. 4From: Measuring the model risk-adjusted performance of machine learning algorithms in credit default predictionBreaking down the models’ riskinessBack to article page