PCNN Fold 0

Back to PCNN overview · Back to main index

Fold metrics

Metric	Value
Test accuracy	11.94%
Test F1 score	0.0889
Hierarchical loss	0.79480463
P-adic loss (total)	1460.53300000
P-adic loss (mean)	0.55176920
Prime base	79
Hidden layer size	27
Max tags	32
Training samples	10,625
Test samples	2,647

P-adic loss breakdown

Agreement	Count	Share	Cost per mistake	Total contribution
Exact match	316	11.94%	0.000000	0.000000
p^3	120	4.53%	0.000002	0.000243
p^2	323	12.20%	0.000160	0.051755
p^1	433	16.36%	0.012658	5.481013
p^0	1,455	54.97%	1.000000	1455.000000

Tag Rank vs First-Layer Weight Magnitude

Tag rank vs max first-layer weight magnitude — Scatter plot showing the relationship between tag battle ranking and maximum absolute first-layer weight value across all hidden units. Shows which input features contribute most to the parameter constrained neural network's hidden representations.

About p-adic loss

P-adic loss measures the distance between predicted and true taxonomy using p-adic metric (base 79). Lower values indicate closer predictions in the taxonomy hierarchy. This metric is shared with the umllr model for comparison.