Table 6 The evaluation indicators of different predictive models in validation set.

From: Combining genetic risk score with artificial neural network to predict the efficacy of folic acid therapy to hyperhomocysteinemia

 

AUC

(95% CI)

Sensitivity(%)

(95% CI)

Specificity (%)

(95% CI)

Youden’s index (95% CI)

Accuracy (%)

(95% CI)

Logistic regression modela

0.878 (0.830–0.925)

76.84 (71.63–81.45)

83.84 (78.32–88.50)

0.6068 (0.5734–0.6358)

80.41 (77.01–83.29)

ANN modelb

0.90 (0.849–0.938)

83.16 (79.63–87.09)

80.81 (76.57–85.29)

0.6397 (0.6051–0.6602)

81.96 (77.24–85.02)

  1. AUC area under the curve, ANN artificial neural network.
  2. aWhen compared with Logistic regression model, there was statistical difference in AUC (P < 0.05).
  3. bWhen compared with ANN model, there was statistical difference in AUC (P < 0.05).
Back to article page