From: Application of three statistical models for predicting the risk of diabetes
Number of first hidden layer nodes | Number of second hidden layer nodes | The area under ROC curve (95%Cl) | Z | P |
---|---|---|---|---|
4 | 1 | 0.724(0.710–0.737) | 18.900 | < 0.0001 |
4 | 2 | 0.736(0.722–0.747) | 18.895 | < 0.0001 |
4 | 3 | 0.745(0.731–0.758) | 20.495 | < 0.0001 |
4 | 4 | 0.748(0.734–0.761) | 19.944 | < 0.0001 |
5 | 1 | 0.752(0.739–0.765) | 19.326 | < 0.0001 |
5 | 2 | 0.760(0.746–0.773) | 23.355 | < 0.0001 |
5 | 3 | 0.753(0.740–0.766) | 19.337 | < 0.0001 |
5 | 4 | 0.750(0.737–0.763) | 20.810 | < 0.0001 |
6 | 1 | 0.741(0.725–0.752) | 19.730 | < 0.0001 |
6 | 2 | 0.742(0.726–0.753) | 20.365 | < 0.0001 |
6 | 3 | 0.740(0.724–0.751) | 19.095 | < 0.0001 |
6 | 4 | 0.737(0.721–0.748) | 18.460 | < 0.0001 |