Fig. 7From: Measuring the model risk-adjusted performance of machine learning algorithms in credit default predictionSHAP interpretation of XGBoost in the empirical exampleBack to article page