Unconstrained Logistic Regression

Model overview

Unconstrained logistic regression classifier using L1 (Lasso) regularization to predict taxonomy IDs from product tags. Unlike parameter-constrained models, this classifier uses ALL available tags and relies on L1 regularization to achieve sparsity by driving unimportant coefficients to zero.

5 CV folds

43.26% Mean accuracy

0.4786 Mean F1

0.28226699 Mean p-adic loss

10,619 Avg non-zero params

Total train samples: 92,388
Total test samples: 23,097
Number of tags (input features): 7,726
Avg non-zero parameters: 10,619 / 7,677,547 (99.9% sparse)
L1 regularization (C): 1.0000

Cross-validation results

Fold	Accuracy	F1	P-adic loss (mean)	Non-zero params	Details
0	43.56%	0.4813	0.28177800	10,517	View
1	42.85%	0.4723	0.28421305	10,651	View
2	43.04%	0.4768	0.28450347	10,772	View
3	43.15%	0.4813	0.29128232	10,511	View
4	43.69%	0.4813	0.26955811	10,642	View