Regression on Synthetic two-feature data Square, n=300
0.988R^2 MeanLightGBM, rel-s features
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LightGBM, rel-s featuresSample size (n)=3002025.12 | 0.988 | 0.0024 | |
| NW + relationsNorm/Embedding Type=L2 norm, Sample size (n)=3002025.12 | 0.9695 | 0.0061 | |
| NW + relationsNorm/Embedding Type=MLP embeddings, Sample size (n)=3002025.12 | 0.9498 | 0.0097 | |
| NW + relationsNorm/Embedding Type=learnable norm, Sample size (n)=3002025.12 | 0.9349 | 0.016 | |
| TabRelSample size (n)=3002025.12 | 0.8597 | 0.1396 | |
| NW, no rel-s infoNorm/Embedding Type=learnable norm, Sample size (n)=3002025.12 | 0.6365 | 0.0551 | |
| NW, no rel-s infoNorm/Embedding Type=L2 norm, Sample size (n)=3002025.12 | 0.618 | 0.0596 | |
| LightGBM, no rel-s infoSample size (n)=3002025.12 | 0.6003 | 0.0702 | |
| NW, rel-s featuresSample size (n)=3002025.12 | -0.926 | 0.2373 |