SSM with Missingness at Random

Coverage

The simulations are based on the make_ssm_data-DGP with \(500\) observations. The simulation considers data under missingness at random.

DoubleML Version                                  0.12.dev0
Script                          SSMMarATECoverageSimulation
Date                                       2025-12-04 21:25
Total Runtime (minutes)                          255.499802
Python Version                                       3.12.3
Config File              scripts/ssm/ssm_mar_ate_config.yml
ⓘLearner g Learner m Learner pi Bias CI Length Coverage
LGBM Regr. LGBM Clas. LGBM Clas. 0.254 1.322 0.973
LGBM Regr. LGBM Clas. Logistic 0.218 1.096 0.980
LGBM Regr. Logistic LGBM Clas. 0.178 0.937 0.970
LassoCV LGBM Clas. LGBM Clas. 0.243 1.271 0.978
LassoCV Logistic Logistic 0.135 0.699 0.966
LassoCV RF Clas. RF Clas. 0.122 0.617 0.957
RF Regr. Logistic RF Clas. 0.140 0.696 0.958
RF Regr. RF Clas. Logistic 0.129 0.668 0.962
RF Regr. RF Clas. RF Clas. 0.123 0.627 0.953
Coverage for 95%-Confidence Interval over 1000 Repetitions
ⓘLearner g Learner m Learner pi Bias CI Length Coverage
LGBM Regr. LGBM Clas. LGBM Clas. 0.254 1.110 0.932
LGBM Regr. LGBM Clas. Logistic 0.218 0.920 0.927
LGBM Regr. Logistic LGBM Clas. 0.178 0.786 0.925
LassoCV LGBM Clas. LGBM Clas. 0.243 1.067 0.937
LassoCV Logistic Logistic 0.135 0.587 0.929
LassoCV RF Clas. RF Clas. 0.122 0.518 0.912
RF Regr. Logistic RF Clas. 0.140 0.584 0.907
RF Regr. RF Clas. Logistic 0.129 0.561 0.917
RF Regr. RF Clas. RF Clas. 0.123 0.526 0.910
Coverage for 90%-Confidence Interval over 1000 Repetitions