Node-level Regression Uncertainty Quantification on ER
100PICPBayesianNN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| BayesianNN2026.05 | 100 | 3.01 | |
| QpiGNNWidth penalty (lambda)=0.12026.05 | 99 | 0.92 | |
| QpiGNNWidth penalty (lambda)=0.52026.05 | 98 | 0.63 | |
| QpiGNNWidth penalty (lambda)=opt.2026.05 | 98 | 0.63 | |
| CF-GNN2026.05 | 90 | 17.15 | |
| RQRadj.-GNN2026.05 | 88 | 0.77 | |
| MC Dropout2026.05 | 76 | 0.23 | |
| SQR-GNN2026.05 | 75 | 0.6 | |
| ER-GNN2026.05 | 72 | 1.86 |