Node-level Regression Uncertainty Quantification on Unemploy.
100PICPER-GNN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ER-GNN2026.05 | 100 | 4.8 | |
| BayesianNN2026.05 | 100 | 2.98 | |
| QpiGNNWidth penalty (lambda)=0.52026.05 | 100 | 0.74 | |
| QpiGNNWidth penalty (lambda)=0.12026.05 | 100 | 0.93 | |
| QpiGNNWidth penalty (lambda)=opt.2026.05 | 100 | 0.73 | |
| RQRadj.-GNN2026.05 | 90 | 0.38 | |
| CF-GNN2026.05 | 90 | 3.16 | |
| CF-GNNoptimized=true2026.05 | 89 | 2.61 | |
| SQR-GNN2026.05 | 87 | 0.33 | |
| MC Dropout2026.05 | 41 | 0.09 |