Response Prediction on PDO tumor-selective (aggregated runs)
68.6AUROCNPF-Flat k = 2
Evaluation results
| Method | Links | ||
|---|---|---|---|
| NPF-Flat k = 2Trainable params=≤ 68,8662026.05 | 68.6 | — | |
| NPF-Tri k = 1Trainable params=≤ 68,8662026.05 | 66.5 | 0.021 | |
| NPF-Gram k = 1Trainable params=≤ 68,8662026.05 | 66.4 | 0.022 | |
| Logistic regression2026.05 | 65.1 | 0.035 | |
| GINTrainable params=18,436–23,3052026.05 | 63.6 | -0.003 | |
| NPF-Tri k = 1Trainable params=≤ 68,8662026.05 | 63.3 | — | |
| NPF-Flat k = 1Trainable params=≤ 68,8662026.05 | 62.9 | 0.057 | |
| GINTrainable params=18,436–23,3052026.05 | 61.8 | 0.068 | |
| PointNetTrainable params=1,465,666–1,481,7352026.05 | 61.1 | 0.075 | |
| PointNet++Trainable params=1,465,666–1,481,7352026.05 | 61.1 | 0.075 | |
| NPF-Gram k = 1Trainable params=≤ 68,8662026.05 | 58.9 | 0.044 | |
| NPF-Flat k = 1Trainable params=≤ 68,8662026.05 | 58.9 | 0.044 | |
| PointTransformerTrainable params=2,154,882–2,165,7672026.05 | 58.7 | 0.099 | |
| NPF-Flat k = 2Trainable params=≤ 68,8662026.05 | 58.1 | 0.052 | |
| GraphSAGETrainable params=36,354–45,4472026.05 | 57.9 | 0.107 | |
| MLPTrainable params=26,562–31,1112026.05 | 57.7 | 0.056 | |
| TDLTrainable params=18,946–23,8152026.05 | 57.7 | 0.056 | |
| GraphSAGETrainable params=36,354–45,4472026.05 | 56.9 | 0.064 | |
| GraphTransformerTrainable params=155,522–173,0632026.05 | 56.5 | 0.068 | |
| MLPTrainable params=26,562–31,1112026.05 | 55.4 | 0.132 | |
| Random forest2026.05 | 55.1 | 0.135 | |
| TDLTrainable params=18,946–23,8152026.05 | 55.1 | 0.135 | |
| Logistic regression2026.05 | 55.1 | 0.082 | |
| GraphTransformerTrainable params=155,522–173,0632026.05 | 54.4 | 0.142 | |
| SVM RBF2026.05 | 52.7 | 0.159 | |
| PointTransformerTrainable params=2,154,882–2,165,7672026.05 | 51.8 | 0.115 | |
| Random forest2026.05 | 51.6 | 0.117 | |
| GCNTrainable params=18,434–23,3032026.05 | 49.3 | 0.193 | |
| GCNTrainable params=18,434–23,3032026.05 | 47.4 | 0.159 | |
| PointNetTrainable params=1,465,666–1,481,7352026.05 | 44.8 | 0.185 | |
| PointNet++Trainable params=1,465,666–1,481,7352026.05 | 44.8 | 0.185 | |
| SVM RBF2026.05 | 41.1 | 0.222 |