Causal Generative Modeling on 11 synthetic datasets Continuous variables (test)
56RF Accuracy (Observed)DBCM
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| DBCMNumber of training samples=10002026.03 | 56 | 31.5023 | 55 | 6.795 | 3.981 | 0.001 | |
| ANMNumber of training samples=10002026.03 | 53 | 8.862 | 54 | 3.022 | 2.622 | 0.5 | |
| KAAMNumber of training samples=10002026.03 | 52 | 16.769 | 53 | 4.765 | 4.662 | 0.055 | |
| CFlowNumber of training samples=10002026.03 | 52 | 20.466 | 61 | 9.618 | 3.861 | 0.001 | |
| KANNumber of training samples=10002026.03 | 51 | 16.498 | 53 | 4.794 | 5.102 | 0.061 |