Bound error estimation on All of Us
-0.01Mean Error (RQ_hat - RQ)Naive IPPW of U~
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Naive IPPW of U~Selection Mechanism=Nonlinear2026.05 | -0.01 | 0.01 | 82 | -0.02 | 0.01 | 11.8 | |
| RakingSelection Mechanism=Nonlinear2026.05 | -0.01 | 0.02 | 79 | -0.06 | 0 | 4.6 | |
| CalibrationSelection Mechanism=Nonlinear2026.05 | -0.01 | 0.02 | 79 | -0.04 | 0 | 3.8 | |
| Ent. BalancingSelection Mechanism=Nonlinear2026.05 | -0.01 | 0.01 | 74 | -0.04 | 0 | 5.3 | |
| UB (true U*)Selection Mechanism=Nonlinear2026.05 | 0.04 | 0.04 | 97 | -0.01 | 0.12 | 28.9 | |
| UB (heuristic U*)Selection Mechanism=Nonlinear2026.05 | 0.06 | 0.05 | 97 | -0.01 | 0.17 | 26.3 |