Back to PCNN overview · Back to main index
| Metric | Value |
|---|---|
| Test accuracy | 11.31% |
| Test F1 score | 0.0854 |
| Hierarchical loss | 0.79245010 |
| P-adic loss (total) | 1473.64560000 |
| P-adic loss (mean) | 0.55947065 |
| Prime base | 79 |
| Hidden layer size | 27 |
| Max tags | 32 |
| Training samples | 10,638 |
| Test samples | 2,634 |
| Agreement | Count | Share | Cost per mistake | Total contribution |
|---|---|---|---|---|
| Exact match | 298 | 11.31% | 0.000000 | 0.000000 |
| p^3 | 111 | 4.21% | 0.000002 | 0.000225 |
| p^2 | 315 | 11.96% | 0.000160 | 0.050473 |
| p^1 | 442 | 16.78% | 0.012658 | 5.594937 |
| p^0 | 1,468 | 55.73% | 1.000000 | 1468.000000 |
P-adic loss measures the distance between predicted and true taxonomy using p-adic metric (base 79). Lower values indicate closer predictions in the taxonomy hierarchy. This metric is shared with the umllr model for comparison.