Back to ULR overview · Back to main index
| Metric | Value |
|---|---|
| Test accuracy | 50.29% |
| Test F1 score | 0.5313 |
| Hierarchical loss | 0.95480896 |
| P-adic loss (total) | 155.47242790 |
| P-adic loss (mean) | 0.22532236 |
| Prime base | 71 |
| Number of tags (input features) | 1,640 |
| Non-zero parameters | 1,551 / 359,379 (99.6% sparse) |
| L1 regularization (C) | 1.0000 |
| Training samples | 2,733 |
| Test samples | 690 |
| Agreement | Count | Share | Cost per mistake | Total contribution |
|---|---|---|---|---|
| Exact match | 347 | 50.29% | 0.000000 | 0.000000 |
| p^6 | 1 | 0.14% | 0.000000 | 0.000000 |
| p^5 | 0 | 0.00% | 0.000000 | 0.000000 |
| p^4 | 10 | 1.45% | 0.000000 | 0.000000 |
| p^3 | 36 | 5.22% | 0.000003 | 0.000101 |
| p^2 | 109 | 15.80% | 0.000198 | 0.021623 |
| p^1 | 32 | 4.64% | 0.014085 | 0.450704 |
| p^0 | 155 | 22.46% | 1.000000 | 155.000000 |
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 (1,551) indicates how many coefficients the model actually uses.