Clinical Outcome Prediction on MIMIC-IV
54.8Mortality (dCorr)MLCI
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| MLCI2026.06 | 54.8 | 49.42 | 51.15 | 61.97 | — | |
| Factorization MachineScore type=FM score2026.06 | 36.41 | 35.82 | 50.8 | 56.77 | — | |
| FusedLogits LR2026.06 | 34.99 | 34.92 | 51.02 | 57.23 | — | |
| Gradient Boosted TreesInput=Score/Logit2026.06 | 34.54 | 34.58 | 51 | 55.61 | — | |
| Multinomial Naive Bayes2026.06 | 32.09 | 32.02 | 35.78 | 45.69 | — | |
| Complement Naive Bayes2026.06 | 32.08 | 32.05 | 35.85 | 45.58 | — | |
| Pure MIL Attention Pooling2026.06 | 29.25 | 29.53 | 48.37 | 52.2 | — | |
| Deep MLP Bag-of-CodesBackbone=EmbeddingBag2026.06 | 28.88 | 29.13 | 48.22 | 52.04 | — | |
| DeepSetsPooling=mean ⊕ max pool2026.06 | 28.51 | 28.84 | 48.6 | 51.92 | — | |
| Set TransformerMode=single-logit2026.06 | 28.5 | 29.02 | 49.12 | 52.04 | — | |
| Deep and Cross Network (DCN)2026.06 | 28.44 | 28.73 | 48.11 | 51.97 | — | |
| Star-GAT2026.06 | 26.97 | 27.48 | 47.19 | 50.24 | — | |
| kNN2026.06 | 22.4 | 23.98 | 30.6 | 34.79 | — | |
| Elixhauser (ECI)2026.06 | 19.98 | 23.87 | 33.15 | 25.98 | — | |
| CCI+ECIFormat=scalar2026.06 | 17.46 | 22.67 | 30.89 | 22.44 | — | |
| Charlson (CCI)2026.06 | 12.59 | 18.44 | 24.55 | 16.09 | — | |
| ProMedStrategy=S#1+GRPO2025.08 | — | — | — | — | 59.9 | |
| ProMedStrategy=S#1+22025.08 | — | — | — | — | 69.5 | |
| Vanilla2025.08 | — | — | — | — | 56.9 |