Training data spans 6,181 products across
369 taxonomies.
Of 24,589 total tags in the dataset,
9,385 tags were used
(tags appearing fewer than 5 times were filtered out).
4,211 products were discarded due to missing or sparse taxonomy labels.Explore the full dataset → | View defective taxonomy labels →
Dummy Baseline
Always predicts most common taxonomy (baseline for comparison)
Arts & Entertainment > Hobbies & Creative Arts > Collectibles
3.2.2
79
1.3%
Tags with strongest signal
Tag
Top taxonomy
Weight
Max |weight|
FRAMED ARTWORK
3.2.2
5.6879
5.6879
WOMENS
1.8.7
5.5699
5.5699
BLUE
14.11.10.4.3
5.4956
5.4956
ACCESSORIES
23.4.4.1.7.4
5.2672
5.2672
GIFT
14.15.1.9
5.1121
5.1121
WHOLESALE
14.11.10.7.9
5.0652
5.0652
VEGAN
13.3.5.2
5.0436
5.0436
KIDS
4.2
5.0249
5.0249
NEW ARRIVALS
13.3.2.8.4
4.8520
4.8520
PLUS SIZE
1.1.1.1.5
4.8185
4.8185
Historical Performance Trends
Tracking model performance and dataset growth over time. Lower p-adic loss indicates better predictions.
Model
Slope (per product)
Intercept
R²
p-value
Importance-Optimised p-adic LR
0.000012
0.2990
0.3100
3.81e-08
PCLR
0.000080
0.2892
0.7407
9.44e-26
PCNN
0.000081
0.2498
0.8422
1.27e-34
ULR
0.000008
0.1789
0.2186
3.19e-04
UNN
0.000026
0.0709
0.7360
2.29e-16
Decision Tree
0.000009
0.1414
0.2947
3.21e-05
Zubarev (UMLLR)
0.000019
0.3064
0.8079
2.81e-16
Zubarev (zeros)
0.000026
0.2990
0.8131
1.60e-16
Zubarev (M1)
0.000007
0.3759
0.3582
2.24e-05
Zubarev (M2)
0.000010
0.3566
0.5340
2.65e-08
Dummy Baseline
-0.000074
1.0973
0.5976
7.09e-15
Extrapolation Analysis: When Will Importance-Optimised p-adic LR Outperform Other Models?
Based on current regression trends, we can extrapolate when Importance-Optimised p-adic LR
will achieve better performance (lower p-adic loss) than other models as the dataset grows.
The confidence intervals are calculated using bootstrap resampling (n=1000).
Statistical Notes: The crossover points are calculated by finding where the
regression lines intersect. The 95% confidence intervals are derived from bootstrap resampling
of the regression parameters. The probability estimates indicate the likelihood that the crossover
will occur given the current trends. Date predictions are based on linear extrapolation of dataset
growth and should be interpreted with caution.
Model
Slope (per tag)
Intercept
R²
p-value
Importance-Optimised p-adic LR
0.000014
0.2422
0.3422
5.13e-09
PCLR
0.000088
-0.0568
0.7266
8.28e-25
PCNN
0.000089
-0.0978
0.8203
2.67e-32
ULR
0.000009
0.1474
0.2519
9.48e-05
UNN
0.000026
-0.0213
0.7613
1.74e-17
Decision Tree
0.000009
0.1085
0.3267
9.61e-06
Zubarev (UMLLR)
0.000020
0.2301
0.8453
3.26e-18
Zubarev (zeros)
0.000027
0.1989
0.8409
5.81e-18
Zubarev (M1)
0.000007
0.3481
0.3701
1.51e-05
Zubarev (M2)
0.000011
0.3180
0.5379
2.22e-08
Dummy Baseline
-0.000079
1.3929
0.6217
8.77e-16
Extrapolation Analysis: When Will Importance-Optimised p-adic LR Outperform Other Models?
Based on current regression trends, we can extrapolate when Importance-Optimised p-adic LR
will achieve better performance (lower p-adic loss) than other models as the dataset grows.
The confidence intervals are calculated using bootstrap resampling (n=1000).
Statistical Notes: The crossover points are calculated by finding where the
regression lines intersect. The 95% confidence intervals are derived from bootstrap resampling
of the regression parameters. The probability estimates indicate the likelihood that the crossover
will occur given the current trends. Date predictions are based on linear extrapolation of dataset
growth and should be interpreted with caution.
Parameter count (log scale) vs p-adic loss. Sparse models use fewer non-zero parameters.
Regression: p-adic loss = slope × log₁₀(params) + intercept
Line
Slope
Intercept
R²
p-value
Significant?
n
With Dummy
-0.0715
0.6486
0.2583
0.1104
No
11
Without Dummy
-0.1316
0.8685
0.1998
0.1953
No
10
Unconstrained models only (no PCLR/PCNN). Both axes on log scale.
