Draft

APO Models

APO Pointwise Coverage

The simulations are based on the the make_irm_data_discrete_treatments-DGP with \(500\) observations. Due to the linearity of the DGP, Lasso and Logit Regression are nearly optimal choices for the nuisance estimation.

DoubleML Version                   0.10.dev0
Script                   irm_apo_coverage.py
Date                     2025-03-05 15:40:07
Total Runtime (seconds)          5746.482011
Python Version                        3.12.9
Table 1: Coverage for 95.0%-Confidence Interval over 1000 Repetitions
Learner g Learner m Treatment Level Bias CI Length Coverage
LGBM LGBM 0.000 2.077 10.316 0.966
LGBM LGBM 1.000 9.216 45.558 0.968
LGBM LGBM 2.000 9.580 44.674 0.952
LGBM Logistic 0.000 1.339 6.704 0.954
LGBM Logistic 1.000 1.693 8.845 0.969
LGBM Logistic 2.000 1.661 8.724 0.969
Linear LGBM 0.000 1.310 6.552 0.950
Linear LGBM 1.000 2.129 12.751 0.983
Linear LGBM 2.000 1.636 8.953 0.967
Linear Logistic 0.000 1.289 6.358 0.953
Linear Logistic 1.000 1.280 6.455 0.957
Linear Logistic 2.000 1.285 6.394 0.957
Table 2: Coverage for 90.0%-Confidence Interval over 1000 Repetitions
Learner g Learner m Treatment Level Bias CI Length Coverage
LGBM LGBM 0.000 2.077 8.658 0.912
LGBM LGBM 1.000 9.216 38.233 0.914
LGBM LGBM 2.000 9.580 37.492 0.892
LGBM Logistic 0.000 1.339 5.626 0.905
LGBM Logistic 1.000 1.693 7.423 0.924
LGBM Logistic 2.000 1.661 7.321 0.920
Linear LGBM 0.000 1.310 5.498 0.900
Linear LGBM 1.000 2.129 10.701 0.948
Linear LGBM 2.000 1.636 7.514 0.933
Linear Logistic 0.000 1.289 5.336 0.903
Linear Logistic 1.000 1.280 5.418 0.909
Linear Logistic 2.000 1.285 5.366 0.906

APOS Coverage

The simulations are based on the the make_irm_data_discrete_treatments-DGP with \(500\) observations. Due to the linearity of the DGP, Lasso and Logit Regression are nearly optimal choices for the nuisance estimation.

The non-uniform results (coverage, ci length and bias) refer to averaged values over all quantiles (point-wise confidende intervals).

DoubleML Version                   0.10.dev0
Script                   irm_apo_coverage.py
Date                     2025-03-05 15:40:07
Total Runtime (seconds)          5746.482011
Python Version                        3.12.9
Table 3: Coverage for 95.0%-Confidence Interval over 1000 Repetitions
Learner g Learner m Bias CI Length Coverage Uniform CI Length Uniform Coverage
LGBM LGBM 7.120 33.615 0.958 40.868 0.973
LGBM Logistic 1.573 8.090 0.962 9.592 0.960
Linear LGBM 1.729 9.418 0.968 11.234 0.974
Linear Logistic 1.283 6.402 0.957 6.818 0.953
Table 4: Coverage for 90.0%-Confidence Interval over 1000 Repetitions
Learner g Learner m Bias CI Length Coverage Uniform CI Length Uniform Coverage
LGBM LGBM 7.120 28.210 0.907 36.194 0.926
LGBM Logistic 1.573 6.789 0.913 8.394 0.922
Linear LGBM 1.729 7.903 0.927 9.850 0.936
Linear Logistic 1.283 5.373 0.905 5.796 0.902

Causal Contrast Coverage

The simulations are based on the the make_irm_data_discrete_treatments-DGP with \(500\) observations. Due to the linearity of the DGP, Lasso and Logit Regression are nearly optimal choices for the nuisance estimation.

The non-uniform results (coverage, ci length and bias) refer to averaged values over all quantiles (point-wise confidende intervals).

DoubleML Version                   0.10.dev0
Script                   irm_apo_coverage.py
Date                     2025-03-05 15:40:07
Total Runtime (seconds)          5746.482011
Python Version                        3.12.9
Table 5: Coverage for 95.0%-Confidence Interval over 1000 Repetitions
Learner g Learner m Bias CI Length Coverage Uniform CI Length Uniform Coverage
LGBM LGBM 9.789 45.131 0.949 51.424 0.965
LGBM Logistic 1.267 6.822 0.964 7.765 0.963
Linear LGBM 1.506 8.855 0.989 10.090 0.992
Linear Logistic 0.294 1.361 0.935 1.550 0.920
Table 6: Coverage for 90.0%-Confidence Interval over 1000 Repetitions
Learner g Learner m Bias CI Length Coverage Uniform CI Length Uniform Coverage
LGBM LGBM 9.789 37.875 0.888 44.829 0.898
LGBM Logistic 1.267 5.725 0.927 6.774 0.927
Linear LGBM 1.506 7.431 0.958 8.799 0.975
Linear Logistic 0.294 1.143 0.874 1.351 0.871