Machine learning-based prediction model of acute kidney injury in patients with acute respiratory distress syndrome

Table 2 Evaluation indicators in testing set of 11 ML models and SOFA score

Model	Accuracy	Sensitivity	Specificity	PPV	NPV	F1_score
Logistic Regression	0.823	0.688	0.851	0.489	0.929	0.571
KNN	0.731	0.500	0.779	0.320	0.882	0.390
Decision Tree	0.753	0.438	0.818	0.333	0.875	0.378
Random Forest	0.850	0.656	0.890	0.553	0.926	0.600
SVM	0.860	0.531	0.929	0.607	0.905	0.567
XGBoost	0.882	0.813	0.896	0.619	0.958	0.703
AdaBoost	0.780	0.375	0.864	0.364	0.869	0.369
GBDT	0.839	0.563	0.896	0.529	0.908	0.545
MLP	0.855	0.594	0.909	0.576	0.915	0.585
LightGBM	0.866	0.625	0.916	0.606	0.922	0.615
CatBoost	0.860	0.531	0.929	0.607	0.905	0.567
SOFA	0.699	0.500	0.740	0.286	0.877	0.364

ML, machine learning; KNN, K-nearest neighbor; SVM, support vector machine; XGBoost, eXtreme gradient boosting; AdaBoost, adaptive boosting; GBDT, gradient boosting decision tree; MLP, multi-layer perception; LightGBM, light gradients boosting machine; CatBoost, category boosting; PPV, positive prediction value; NPV, negative prediction value; SOFA, sequential organ failure assessment

ISSN: 1471-2466