Table 10 AUC for heart disease dataset.
From: Using item response theory as a methodology to impute categorical missing values
Type | Missing (%) | KNN | MICE\(^1\) ± Std. err. | Datawig\(^1\) ± Std. err. | IRT |
|---|---|---|---|---|---|
MAR | 2389 | 0.832 | 0.831 ± 0.0001 | 0.831 ± 0.0001 | 0.830 |
MAR | 4779 | 0.830 | 0.830 ± 0.0002 | 0.831 ± 0.0002 | 0.832 |
MAR | 14,336 | 0.827 | 0.828 ± 0.0001 | 0.829 ± 0.0003 | 0.827 |
MAR | 23,893 | 0.827 | 0.827 ± 0.0003 | 0.828 ± 0.0002 | 0.829 |
Marginal means | 0.829 | 0.829 | 0.830 | 0.829 | |
MCAR | 2389 | 0.829 | 0.831 ± 0.0003 | 0.832 ± 0.0004 | 0.832 |
MCAR | 4779 | 0.831 | 0.830 ± 0.0005 | 0.831 ± 0.0002 | 0.831 |
MCAR | 14,336 | 0.829 | 0.828 ± 0.0005 | 0.829 ± 0.0002 | 0.828 |
MCAR | 23,893 | 0.828 | 0.826 ± 0.0003 | 0.828 ± 0.0002 | 0.829 |
Marginal means | 0.829 | 0.829 | 0.830 | 0.830 | |
Original data AUC: | 0.8313 | Â | |||
