Imputation on Semi-synthetic EHR dataset Pooled 30-80% MNAR (summary)
2.38Mean RankGPT-5.4
Evaluation results
| Method | Links | |
|---|---|---|
| GPT-5.4framework=LLM-driven evolutionary imputation2026.05 | 2.38 | |
| GPT-OSS-120bframework=LLM-driven evolutionary imputation2026.05 | 2.75 | |
| Qwen3.5-Plusframework=LLM-driven evolutionary imputation2026.05 | 3.38 | |
| MissForest2026.05 | 3.75 | |
| CausalCFM2026.05 | 3.92 | |
| LOCF2026.05 | 4.83 |