Causal Discovery on Synthetic Scale-Free networks d=15 SF
11SHDDCD-PFN
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| DCD-PFNdimension (d)=15, zero-shot=true, number of samples (n)=8002026.06 | 11 | 88.3 | 26 | 39.1 | |
| DirectLiNGAMdimension (d)=15, number of samples (n)=8002026.06 | 11.36 | 95 | 19.1 | 31.3 | |
| GOLEMdimension (d)=15, number of samples (n)=8002026.06 | 11.68 | 99 | 16.9 | 28.1 | |
| ICALiNGAMdimension (d)=15, number of samples (n)=8002026.06 | 11.92 | 87 | 15.1 | 25.3 | |
| AVICIdimension (d)=15, zero-shot=true, number of samples (n)=8002026.06 | 12.56 | 69.1 | 18.6 | 28.1 | |
| Notearsdimension (d)=15, number of samples (n)=8002026.06 | 12.64 | 78.7 | 10 | 17.2 | |
| PC (E-KCIT)dimension (d)=15, number of samples (n)=8002026.06 | 12.88 | 16.9 | 11.4 | 13.5 | |
| GESdimension (d)=15, number of samples (n)=8002026.06 | 13.708 | 20.5 | 16.1 | 17.8 | |
| PC (KCIT)dimension (d)=15, number of samples (n)=8002026.06 | 14.4 | 10.6 | 9.4 | 10 | |
| GraNDAGdimension (d)=15, number of samples (n)=8002026.06 | 19.68 | 38.6 | 16.6 | 18.7 |