ULR Fold 0

Back to ULR overview · Back to main index

Fold metrics

MetricValue
Test accuracy50.92%
Test F1 score0.5467
Hierarchical loss0.90454930
P-adic loss (total)400.33663966
P-adic loss (mean)0.26165793
Prime base79
Number of tags (input features)5,904
Non-zero parameters4,328 / 3,253,655 (99.9% sparse)
L1 regularization (C)1.0000
Training samples6,160
Test samples1,530

P-adic loss breakdown

AgreementCountShareCost per mistakeTotal contribution
Exact match78151.05%0.0000000.000000
p^630.20%0.0000000.000000
p^540.26%0.0000000.000000
p^4382.48%0.0000000.000001
p^3764.97%0.0000020.000154
p^21258.17%0.0001600.020029
p^11046.80%0.0126581.316456
p^039926.08%1.000000399.000000

About L1 regularization

L1 (Lasso) regularization promotes sparsity by driving many coefficients to exactly zero. This model uses ALL available tags (5,904) but L1 regularization selects which features are actually used. The number of non-zero parameters (4,328) indicates how many coefficients the model actually uses.