From: Interpretable instance disease prediction based on causal feature selection and effect analysis
Dataset | XGBOOST | With LASSO | Factual prediction network | Counterfactual prediction network |
---|---|---|---|---|
AUROC | Â | Â | Â | Â |
 Dataset1 | .574 ± 0.10 | .498 ± 0.06 | .681 ± 0.02 | .693 ± 0.06 |
 Dataset2 | .872 ± 0.03 | .823 ± 0.61 | .864 ± 0.61 | .877 ± 0.03 |
 Dataset3 | .899 ± 0.01 | .862 ± 0.03 | .890 ± 0.03 | .911 ± 0.02 |
AUPRC | Â | Â | Â | Â |
 Dataset1 | .577 ± 1.02 | .499 ± 0.08 | .681 ± 0.04 | .694 ± 0.03 |
 Dataset2 | .878 ± 0.31 | .591 ± 0.37 | .861 ± 0.21 | .886 ± 0.04 |
 Dataset3 | .904 ± 0.04 | .890 ± 0.02 | .890 ± 0.05 | .905 ± 0.02 |