Back to ULR overview · Back to main index
| Metric | Value |
|---|---|
| Test accuracy | 51.64% |
| Test F1 score | 0.5557 |
| Hierarchical loss | 0.90522592 |
| P-adic loss (total) | 431.43967993 |
| P-adic loss (mean) | 0.24035637 |
| Prime base | 79 |
| Number of tags (input features) | 3,664 |
| Non-zero parameters | 4,541 / 1,792,185 (99.7% sparse) |
| L1 regularization (C) | 1.0000 |
| Training samples | 7,315 |
| Test samples | 1,795 |
| Agreement | Count | Share | Cost per mistake | Total contribution |
|---|---|---|---|---|
| Exact match | 927 | 51.64% | 0.000000 | 0.000000 |
| p^6 | 2 | 0.11% | 0.000000 | 0.000000 |
| p^5 | 0 | 0.00% | 0.000000 | 0.000000 |
| p^4 | 28 | 1.56% | 0.000000 | 0.000001 |
| p^3 | 82 | 4.57% | 0.000002 | 0.000166 |
| p^2 | 215 | 11.98% | 0.000160 | 0.034450 |
| p^1 | 111 | 6.18% | 0.012658 | 1.405063 |
| p^0 | 430 | 23.96% | 1.000000 | 430.000000 |
L1 (Lasso) regularization promotes sparsity by driving many coefficients to exactly zero. This model uses ALL available tags (3,664) but L1 regularization selects which features are actually used. The number of non-zero parameters (4,541) indicates how many coefficients the model actually uses.