Table 3 The comparison of causal variable determination for the toy datasets between ICPs and the proposed method
From: Automated causal inference in application to randomized controlled clinical trials
Finite sample settinga | |||
|---|---|---|---|
JS (FWER) | |||
Two confounders | One confounder | Zero confounders | |
ICP | 0.332 (0.98) | 0.382 (0.85) | 1.00 (0.00) |
NICP | 0.333 (0.98) | 0.384 (0.85) | 1.00 (0.00) |
AICP | 0.439 (0.05) | 0.483 (0.11) | 0.998 (0.0004) |
Proposed | 0.911 (0.13) | 0.923 (0.14) | 0.994 (0.006) |
ABCD settingb | |||
|---|---|---|---|
JS (FWER) | |||
Two confounders | One confounder | Zero confounders | |
ICP | 0.517 (0.61) | 0.559 (0.51) | 0.976 (0.00) |
NICP | 0.512 (0.69) | 0.558 (0.57) | 0.992 (0.002) |
AICP | 0.417 (0.12) | 0.437 (0.18) | 0.991 (0.01) |
Proposed | 0.922 (0.08) | 0.928 (0.12) | 0.985 (0.02) |
