Table 3 Prediction results for the different models on MIMIC-IIItest dataset, where the best AUROC and AUPRC are in bold
From: CRISP: A causal relationships-guided deep learning framework for advanced ICU mortality prediction
Type | Models | AUROC | AUPRC |
|---|---|---|---|
Score-based | SAPS II | 0.805 | 0.337 |
Time series | GRU-D | 0.835 | 0.359 |
Time series | IPNET | 0.853 | 0.418 |
Time series | MC | 0.870 | 0.533 |
Time series | MTRNN | 0.895 | 0.520 |
Text | IHM-AS | 0.918 | 0.625 |
Hybrid | MultiModal - 1DCNN | 0.865 | 0.525 |
Hybrid | Vital + EntityEmb | 0.874 | 0.529 |
Hybrid | DECAF | 0.893 | - |
Hybrid | GAN (c-med GAN) | 0.910 | 0.532 |
Hybrid | MMDL | 0.940 | 0.772 |
Hybrid | CRISP | 0.948 | 0.761 |