Back to ULR overview · Back to main index
| Metric | Value |
|---|---|
| Test accuracy | 43.69% |
| Test F1 score | 0.4813 |
| Hierarchical loss | 0.89140898 |
| P-adic loss (total) | 1249.94097469 |
| P-adic loss (mean) | 0.26955811 |
| Prime base | 83 |
| Number of tags (input features) | 7,726 |
| Non-zero parameters | 10,642 / 7,703,819 (99.9% sparse) |
| L1 regularization (C) | 1.0000 |
| Training samples | 18,460 |
| Test samples | 4,637 |
| Agreement | Count | Share | Cost per mistake | Total contribution |
|---|---|---|---|---|
| Exact match | 2,026 | 43.69% | 0.000000 | 0.000000 |
| p^6 | 2 | 0.04% | 0.000000 | 0.000000 |
| p^5 | 5 | 0.11% | 0.000000 | 0.000000 |
| p^4 | 91 | 1.96% | 0.000000 | 0.000002 |
| p^3 | 279 | 6.02% | 0.000002 | 0.000488 |
| p^2 | 586 | 12.64% | 0.000145 | 0.085063 |
| p^1 | 403 | 8.69% | 0.012048 | 4.855422 |
| p^0 | 1,245 | 26.85% | 1.000000 | 1245.000000 |
L1 (Lasso) regularization promotes sparsity by driving many coefficients to exactly zero. This model uses ALL available tags (7,726) but L1 regularization selects which features are actually used. The number of non-zero parameters (10,642) indicates how many coefficients the model actually uses.