Machine learning models for predicting acute kidney injury after pediatric living donor liver transplantation in biliary atresia

Table 2 Summary of model performance comparison

Model	Accuracy (mean)	Accuracy (95%CI)	Precision (mean)	Precision (95%CI)	Recall (mean)	Recall (95%CI)	F1 score (mean)	F1 score (95%CI)	AUC (mean)	AUC (95%CI)
Random forest	0.842	0.783-0.902	0.805	0.663-0.947	0.599	0.478-0.720	0.678	0.563-0.793	0.875	0.805-0.944
XGBoost	0.806	0.740-0.872	0.685	0.560-0.810	0.634	0.484-0.785	0.641	0.532-0.751	0.848	0.771-0.925
Gradient boosting	0.809	0.759-0.859	0.706	0.585-0.827	0.611	0.481-0.741	0.636	0.547-0.726	0.841	0.764-0.917
AdaBoost	0.818	0.758-0.879	0.74	0.594-0.885	0.544	0.408-0.681	0.619	0.489-0.750	0.824	0.730-0.917
KNN	0.782	0.739-0.825	0.677	0.544-0.810	0.468	0.344-0.592	0.533	0.424-0.641	0.823	0.766-0.880
MLP	0.8	0.763-0.837	0.661	0.587-0.735	0.612	0.476-0.749	0.62	0.540-0.701	0.816	0.739-0.892
SVM	0.812	0.766-0.859	0.748	0.624-0.873	0.521	0.396-0.647	0.598	0.492-0.705	0.81	0.748-0.872
Logistic regression	0.788	0.740-0.836	0.689	0.535-0.843	0.444	0.334-0.555	0.533	0.416-0.651	0.79	0.708-0.873
Decision tree	0.779	0.712-0.846	0.638	0.498-0.778	0.631	0.521-0.741	0.619	0.524-0.714	0.734	0.664-0.803

KNN: K-nearest neighbors; MLP: Multilayer perceptron; SVM: Support vector machine; AUC: Area under the curve; CI: Confidence interval.