Parameter Constrained Logistic Regression

Predicting Shopify taxonomy from product tags using parameter constrained logistic regression

← Back to index

Model summary

This parameter constrained logistic regression model predicts taxonomy IDs from product tags. Trained at: 2026-04-14T05:31:59+10:00

5CV folds

6.59%Mean test accuracy

557Taxonomy classes

Aggregate statistics

Total train samples: 43,740
Total test samples: 10,935
Mean F1 (weighted): 0.0614
Mean p-adic loss: 0.75835364

Cross-validation fold results

Average p-adic loss across all folds: 0.75835364

Fold	Test Accuracy	F1 Score	P-adic Loss (mean)	Details
0	6.38%	0.0639	0.81229926	View details →
1	7.20%	0.0664	0.74664182	View details →
2	6.76%	0.0608	0.79080885	View details →
3	6.51%	0.0633	0.72299122	View details →
4	6.08%	0.0524	0.71902708	View details →

Largest taxonomy classes

Taxonomy ID	Path	Samples	Share
gid://shopify/TaxonomyCategory/bt	4	106	0.97%
gid://shopify/TaxonomyCategory/lb	15	88	0.80%
gid://shopify/TaxonomyCategory/bu	5	47	0.43%
gid://shopify/TaxonomyCategory/pa	19	20	0.18%
gid://shopify/TaxonomyCategory/na	25	18	0.16%
gid://shopify/TaxonomyCategory/sg	23	13	0.12%
gid://shopify/TaxonomyCategory/os	18	12	0.11%
gid://shopify/TaxonomyCategory/gc	11	11	0.10%
gid://shopify/TaxonomyCategory/hg	14	8	0.07%
gid://shopify/TaxonomyCategory/el	8	7	0.06%
gid://shopify/TaxonomyCategory/ma	16	7	0.06%
gid://shopify/TaxonomyCategory/fb	9	6	0.05%
gid://shopify/TaxonomyCategory/me	17	6	0.05%
gid://shopify/TaxonomyCategory/bi	6	4	0.04%

Tags with strongest signal

Tag	Taxonomy ID	Path	Weight	Max \|weight\|
No tag signals available