Table 3.

Performance of four models for predicting the occurrence of cervical cancer.

Model	AUC		Accuracy		Sensitivity		Specificity
	Training	Validation	Training	Validation	Training	Validation	Training	Validation
Logistic regression	0.813	0.772	0.712	0.688	0.907	0.870	0.616	0.598
Supportive vector machine	0.752	0.639	0.745	0.645	0.678	0.600	0.822	0.679
Random forest	0.671	0.703	0.663	0.696	0.745	0.778	0.597	0.627
Decisive tree	0.718	0.726	0.718	0.725	0.763	0.775	0.669	0.672
eXtreme gradient boosting	0.807	0.768	0.741	0.703	0.911	0.836	0.606	0.597
Neural network	0.812	0.772	0.712	0.688	0.931	0.902	0.536	0.519