From: DeepMPM: a mortality risk prediction model using longitudinal EHR data
Model | AUC | Precision | Recall | F1-score |
---|---|---|---|---|
RNN | 0.8318 ± 0.0102 | 0.7392 ± 0.0340 | 0.7571 ± 0.0408 | 0.7505 ± 0.0139 |
Multi-task learning | 0.6330 ± 0.0084 | 0.6130 ± 0.0245 | 0.5808 ± 0.1439 | 0.5868 ± 0.0674 |
LSTM-NN | 0.8326 ± 0.0087 | 0.7562 ± 0.0289 | 0.7508 ± 0.0519 | 0.7621 ± 0.0148 |
RETAIN | 0.8268 ± 0.0081 | 0.7592 ± 0.0103 | 0.7788 ± 0.0091 | 0.7687 ± 0.0089 |
Deepcare | 0.7876 ± 0.0098 | 0.7858 ± 0.0264 | 0.7707 ± 0.0458 | 0.7782 ± 0.0147 |
DeepMPM-w/o-\(\beta\) | 0.8435 ± 0.0073 | 0.7685 ± 0.0210 | 0.7759 ± 0.0490 | 0.7710 ± 0.0177 |
DeepMPM | 0.8501 ± 0.0076 | 0.7700 ± 0.0306 | 0.7987 ± 0.0538 | 0.7824 ± 0.0153 |