Postoperative complication prediction on POC-L (test)
73.83TPR @ FPR=0.2Multivariate logistic regression
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Multivariate logistic regression2026.01 | 73.83 | 80.89 | 80.37 | |
| MIRACLEInput data modality=Multimodal, Backbone LLM=DeepSeek R1-Distill Qwen-32B2026.01 | 73.83 | 80.94 | 81.31 | |
| MIRACLEInput data modality=Multimodal, Backbone LLM=OpenBioLLM-70B2026.01 | 71.96 | 81.04 | 81.31 | |
| MIRACLEInput data modality=Multimodal, Backbone LLM=Llama 3.3 70B-Instruct2026.01 | 71.03 | 80.84 | 81.31 | |
| Gradient Boosting Classifier2026.01 | 65.42 | 78.53 | 67.29 | |
| Random Forest Classifier2026.01 | 62.62 | 77 | 74.76 | |
| DeepSeek R1-Distill Qwen-32BInput data modality=Clinical-only2026.01 | 54.21 | 64.49 | 56.07 | |
| XGBoost2026.01 | 53.27 | 75.17 | 64.48 | |
| OpenBioLLM-70BInput data modality=Clinical-only2026.01 | 52.34 | 71.01 | 60.75 | |
| LightGBM2026.01 | 46.73 | 74.77 | 69.16 | |
| Surgeons2026.01 | 44.86 | — | — | |
| Llama 3.3 70B-InstructInput data modality=Clinical-only2026.01 | 41.12 | 69.68 | 74.77 |