Response Prediction on PDO EFP (aggregated aggregate runs)
0.71AUROCGraphTransformer
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GraphTransformerTrainable params=155,522–173,0632026.05 | 0.71 | -0.004 | |
| NPF-Flat k = 1Trainable params=≤ 68,8662026.05 | 0.706 | — | |
| NPF-Flat k = 2Trainable params=≤ 68,8662026.05 | 0.704 | 0.002 | |
| NPF-Tri k = 1Trainable params=≤ 68,8662026.05 | 0.7 | 0.006 | |
| GINTrainable params=18,436–23,3052026.05 | 0.685 | 0.021 | |
| NPF-Gram k = 1Trainable params=≤ 68,8662026.05 | 0.664 | 0.042 | |
| MLPTrainable params=26,562–31,1112026.05 | 0.652 | 0.054 | |
| GraphSAGETrainable params=36,354–45,4472026.05 | 0.627 | 0.079 | |
| Logistic regression2026.05 | 0.615 | 0.091 | |
| SVM RBF2026.05 | 0.563 | 0.143 | |
| PointTransformerTrainable params=2,154,882–2,165,7672026.05 | 0.558 | 0.148 | |
| TDLTrainable params=18,946–23,8152026.05 | 0.549 | 0.157 | |
| GCNTrainable params=18,434–23,3032026.05 | 0.534 | 0.172 | |
| Random forest2026.05 | 0.501 | 0.205 | |
| PointNetTrainable params=1,465,666–1,481,7352026.05 | 0.494 | 0.212 | |
| PointNet++Trainable params=1,465,666–1,481,7352026.05 | 0.494 | 0.212 |