TABLE 2.

Gestational diabetes mellitus prediction models in early pregnancy.

Model	Random forest model					Logistic regression model
	AUC (95% CI)	Accuracy	Recall	Precision	F1 score	AUC (95% CI)	Accuracy	Sensitivity	Specificity
Model 1	0.742 (0.528, 0.912)	0.718	0.667	0.778	0.718	0.772 (0.663, 0.891)	0.751	0.832	0.672
Model 2	0.733 (0.515, 0.902)	0.763	0.789	0.750	0.769	0.881 (0.801, 0.963)	0.822	0.889	0.743
Model 3	0.769 (0.473, 0.931)	0.774	0.846	0.688	0.759	0.849 (0.772, 0.942)	0.772	0.781	0.769
Model 4	0.893 (0.736, 0.990)	0.878	0.941	0.800	0.865	0.930 (0.872, 0.991)	0.872	0.861	0.891

It shows the results in the training dataset; GDM, gestational diabetes mellitus; AUC, area under curve; CI, confidence interval.