Table 2 The performance of machine learning models for stillbirth prediction in the test set for the whole population. AUC area under the curve, AUPR area under precision-recall curve.
From: Prediction of stillbirth using machine learning methods
Stillbirth type | Period | AUC | AUPR | Specificity | F1 score | Precision | Recall |
|---|---|---|---|---|---|---|---|
All stillbirth | Baseline | 0.720 (0.711–0.729) | 0.016 (0.014–0.019) | 0.732 (0.723–0.741) | 0.040 (0.036–0.044) | 0.021 (0.018–0.023) | 0.611 (0.602–0.621) |
E1 | 0.740 (0.731–0.748) | 0.019 (0.017–0.022) | 0.850 (0.843–0.857) | 0.048 (0.043–0.052) | 0.025 (0.022–0.028) | 0.422 (0.412–0.432) | |
Early stillbirth | Baseline | 0.728 (0.719–0.737) | 0.010 (0.008–0.011) | 0.958 (0.954–0.962) | 0.030 (0.027–0.033) | 0.017 (0.014–0.019) | 0.146 (0.139–0.153) |
E1 | 0.743 (0.734–0.751) | 0.094 (0.089-0.100) | 0.902 (0.896–0.908) | 0.029 (0.026–0.033) | 0.015 (0.013–0.018) | 0.313 (0.303–0.322) | |
Late stillbirth | Baseline | 0.705 (0.696–0.714) | 0.011 (0.009–0.013) | 0.998 (0.997–0.999) | 0.062 (0.057–0.066) | 0.087 (0.081–0.093) | 0.048 (0.043–0.052) |
E1 | 0.719 (0.710–0.728) | 0.024 (0.021–0.027) | 0.977 (0.974–0.980) | 0.036 (0.032–0.040) | 0.021 (0.018–0.024) | 0.119 (0.113–0.125) | |
T0 | 0.781 (0.773–0.789) | 0.015 (0.013–0.017) | 0.967 (0.963–0.970) | 0.053 (0.048–0.057) | 0.030 (0.026–0.033) | 0.238 (0.230–0.247) |
