Artificial neural network models for prediction of cardiovascular autonomic dysfunction in general Chinese population

Table 3 Prediction models using artificial neural network

	Model 1	Model 2	Model 3	Model 4	Model 5	Mean ± SD	95% CI
AUC	0.738	0.763	0.737	0.783	0.789	0.762 ± 0.025	0.732–0.793
Cutoff	0.234	0.229	0.216	0.227	0.175	0.216 ± 0.024	0.187–0.246
Sensitivity	0.694	0.789	0.677	0.777	0.821	0.751 ± 0.065	0.667–0.828
Specificity	0.694	0.663	0.647	0.704	0.618	0.665 ± 0.035	0.622–0.709
Yuden Index	0.388	0.452	0.324	0.481	0.439	0.413 ± 0.063	0.334–0.491
PPV	0.332	0.339	0.301	0.373	0.321	0.330 ± 0.026	0.298–0.361
NPV	0.912	0.935	0.898	0.932	0.94	0.924 ± 0.018	0.902–0.945
HL Statistics	14.64	8.143	8.421	7.424	7.196	9.165 ± 3.103	5.313–13.017
Accuracy	0.695	0.685	0.651	0.714	0.661	0.681 ± 0.026	0.650–0.713

Note: AUC: area under the receiver-operating curve, PPV: positive predictive value; NPV: negative predictive value.

ISSN: 1472-6947