Table 2.

Performance metrics of the models for prediction of postoperative AKI/ Severe AKI in the deviation set

Outcome	Model	Sensitivity (95%CI)	Specificity (95%CI)	AUC# (95%CI)	Acurracy (95%CI)	Youden index (95%CI)
AKI	LR	0.774(0.719,0.813)	0.739(0.698,0.784)	0.812(0.756,0.860)	0.753(0.719,0.781)	0.513(0.451,0.573)
	DT	0.473(0.420,0.537)	0.594(0.544,0.635)	0.534(0.467,0.599)	0.545(0.516,0.570)	0.067(-0.010,0.194)
	RF	0.602(0.552,0.657)	0.725(0.673,0.777)	0.712(0.649,0.770)	0.675(0.626,0.709)	0.327(0.238,0.379)
	GBC	0.581(0.525,0.629)	0.775(0.725,0.814)	0.732(0.670,0.788)	0.697(0.672,0.742)	0.356(0.294,0.436)
	GNB	0.452(0.358,0.507)	0.826(0.781,0.859)	0.762(0.701,0.815)	0.675(0.644,0.702)	0.278(0.205,0.347)
	MLP	0.656(0.575,0.729)	0.804(0.749,0.843)	0.793(0.735,0.844)	0.745(0.705,0.778)	0.460(0.391,0.536)
Severe AKI	LR	0.148(0.078,0.250)	0.990(0.982,0.997)	0.803(0.746,0.852)	0.892(0.877,0.905)	0.138(0.065,0.260)
	DT	0.593(0.483,0.707)	0.907(0.882,0.930)	0.749(0.688,0.803)	0.870(0.841,0.894)	0.500(0.393,0.596)
	RF	0.296(0.181,0.388)	0.975(0.964,0.985)	0.805(0.748,0.854)	0.896(0.875,0.912)	0.271(0.147,0.348)
	GBC	0.333(0.224,0.431)	0.971(0.955,0.983)	0.86(0.808,0.902)	0.896(0.868,0.917)	0.304(0.202,0.424)
	GNB	0.407(0.284,0.517)	0.922(0.894,0.942)	0.734(0.672,0.790)	0.861(0.835,0.886)	0.329(0.219,0.428)
	MLP	0.074(0.009,0.138)	1.000(1.000,1.000)	0.718(0.655,0.775)	0.892(0.888,0.898)	0.074(0.026,0.147)

Abbreviations: AKI, acute kidney injury; AUC, area under receiver operating characteristic curves; CI, confidence interval; LR, Logistic Regression; DT, Decision Tree; RF, Random Forest; GBC, Gradient Boosting Classifier; GNB, Gaussian Naive Bayes; MLP, Multilayer Perceptron;