Table 2.

Performance parameters of diabetes prediction models.

		AUROC	Sensitivity	Specificity	PPV	NPV	Accuracy
Logistic regression	MI by Median	0.547	0.964	0.529	0.996	0.098	0.960
Decision tree	MI by Median	0.590	0.967	0.582	0.994	0.186	0.962
	No MI	0.598	0.968	0.556	0.993	0.204	0.962
Random Forest	MI by Median	0.524	0.962	0.696	0.999	0.050	0.961
	No MI	0.530	0.962	0.756	0.999	0.061	0.962
XGBoost	MI by Median	0.616	0.969	0.646	0.994	0.237	0.964
	No MI	0.623	0.970	0.690	0.995	0.250	0.966

AUROC—area under the receiver operating characteristic curve; NPV—negative prediction value; PPV—positive prediction value; MI—missing value imputation.