Back to ULR overview · Back to main index
| Metric | Value |
|---|---|
| Test accuracy | 54.49% |
| Test F1 score | 0.5829 |
| Hierarchical loss | 0.95862737 |
| P-adic loss (total) | 141.57854366 |
| P-adic loss (mean) | 0.21549246 |
| Prime base | 71 |
| Number of tags (input features) | 1,640 |
| Non-zero parameters | 1,650 / 357,738 (99.5% sparse) |
| L1 regularization (C) | 1.0000 |
| Training samples | 2,766 |
| Test samples | 657 |
| Agreement | Count | Share | Cost per mistake | Total contribution |
|---|---|---|---|---|
| Exact match | 358 | 54.49% | 0.000000 | 0.000000 |
| p^6 | 1 | 0.15% | 0.000000 | 0.000000 |
| p^5 | 0 | 0.00% | 0.000000 | 0.000000 |
| p^4 | 10 | 1.52% | 0.000000 | 0.000000 |
| p^3 | 31 | 4.72% | 0.000003 | 0.000087 |
| p^2 | 76 | 11.57% | 0.000198 | 0.015076 |
| p^1 | 40 | 6.09% | 0.014085 | 0.563380 |
| p^0 | 141 | 21.46% | 1.000000 | 141.000000 |
L1 (Lasso) regularization promotes sparsity by driving many coefficients to exactly zero. This model uses ALL available tags (1,640) but L1 regularization selects which features are actually used. The number of non-zero parameters (1,650) indicates how many coefficients the model actually uses.