PCNN Fold 3

Back to PCNN overview · Back to main index

Fold metrics

Metric	Value
Test accuracy	13.45%
Test F1 score	0.0908
Hierarchical loss	0.77097195
P-adic loss (total)	1147.88610000
P-adic loss (mean)	0.74586490
Prime base	79
Hidden layer size	27
Max tags	32
Training samples	6,151
Test samples	1,539

P-adic loss breakdown

Agreement	Count	Share	Cost per mistake	Total contribution
Exact match	207	13.45%	0.000000	0.000000
p^4	1	0.06%	0.000000	0.000000
p^3	36	2.34%	0.000002	0.000073
p^2	79	5.13%	0.000160	0.012658
p^1	69	4.48%	0.012658	0.873418
p^0	1,147	74.53%	1.000000	1147.000000

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.