Fig. 4From: Machine learning-based prediction model of acute kidney injury in patients with acute respiratory distress syndromea SHAP values output by all patients in the XGBoost model; b The feature importance of XGBoost model. SHAP, shapley additive explanations; XGBoost, eXtreme gradient boosting; PO2, partial pressure of carbon dioxide; BUN, blood urea nitrogen; WBC, white blood cell; UO, urine output; ALB, albumin; AST, aspartate aminotransferase; SpO2, oxygen saturation; ALT, alanine aminotransferase; DBP, diastolic blood pressure; RBC, red blood cell; TBIL, total bilirubin; RDW, red cell volume distribution width; HR, heart rate; BE, base excess; UTI, urinary tract infectionBack to article page