Fig. 3From: Measuring the model risk-adjusted performance of machine learning algorithms in credit default predictionComparison between different model purposesBack to article page