Table 4 Summary of model performance in different scenarios
From: Predicting interval from diagnosis to delivery in preeclampsia using electronic health records
| Â | Mean C-index | AUROC of delivery prediction on UM test set | AUROC of delivery prediction on UF test set | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Â | train | test | validation | deliver in 2 days | deliver in 7 days | deliver in 14 days | deliver in 2 days | deliver in 7 days | deliver in 14 days |
baseline model | 0.73 | 0.74 | 0.73 | 0.85 | 0.88 | 0.89 | 0.68 | 0.77 | 0.78 |
full model | 0.78 | 0.79 | 0.74 | 0.88 | 0.93 | 0.93 | 0.84 | 0.9 | 0.91 |
baseline model on eope patients | 0.67 | 0.68 | 0.63 | 0.64 | 0.76 | 0.84 | 0.63 | 0.67 | 0.7 |
full model on eope patients | 0.74 | 0.76 | 0.67 | 0.87 | 0.82 | 0.82 | 0.71 | 0.72 | 0.71 |
