PCNN Fold 4

Back to PCNN overview · Back to main index

Fold metrics

Metric	Value
Test accuracy	16.28%
Test F1 score	0.1186
Hierarchical loss	0.92388840
P-adic loss (total)	812.10370000
P-adic loss (mean)	0.65439457
Prime base	71
Hidden layer size	27
Max tags	32
Training samples	4,940
Test samples	1,241

P-adic loss breakdown

Agreement	Count	Share	Cost per mistake	Total contribution
Exact match	202	16.28%	0.000000	0.000000
p^4	4	0.32%	0.000000	0.000000
p^3	51	4.11%	0.000003	0.000142
p^2	96	7.74%	0.000198	0.019044
p^1	77	6.20%	0.014085	1.084507
p^0	811	65.35%	1.000000	811.000000

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 71). Lower values indicate closer predictions in the taxonomy hierarchy. This metric is shared with the umllr model for comparison.