Table 3 Comparison of performance indices of the artificial neural network (ANN), MLR, and Cox regression models for predicting 5-year postoperative mortality of breast cancer patients
From: Predictive model for 5-year mortality after breast cancer surgery in Taiwan residents
Dataset | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | Accuracy (%) | AUC (%) |
|---|---|---|---|---|---|---|
Training dataset (n = 2543) | ||||||
ANN model | 86.88 | 89.09 | 65.24 | 87.91 | 87.43 | 72.86 |
MLR model | 83.72 | 87.73 | 64.80 | 83.57 | 86.55 | 52.42 |
Cox model | 86.37 | 88.06 | 60.24 | 85.48 | 84.53 | 65.13 |
Testing dataset (n = 726) | ||||||
ANN model | 88.00 | 89.04 | 77.52 | 87.64 | 88.50 | 71.67 |
MLR model | 83.00 | 84.36 | 73.35 | 87.02 | 86.09 | 51.93 |
Cox model | 86.43 | 87.57 | 74.99 | 87.10 | 86.45 | 65.87 |
Validation dataset (n = 363) | ||||||
ANN model | 85.66 | 88.71 | 53.42 | 86.89 | 88.82 | 70.76 |
MLR model | 83.77 | 87.35 | 50.96 | 86.01 | 85.40 | 52.44 |
Cox model | 84.61 | 85.43 | 52.02 | 86.60 | 86.91 | 64.79 |