PDP report
target y · 4 features · 3 plotted
1 · Overview — where to look
An additive surrogate read off the global curves reproduces 84.9% of the model's predicted variance; adding the regional plots kept by the decision sequence, 99.8%.
Each point is a feature: importance (x) against heterogeneity (y). Bottom-left is ignorable; bottom-right is important and fully described by its mean effect; the top-right corner — important and heterogeneous — is where the mean hides something. An arrow marks each split the decision sequence accepted: from the feature's global point to its weighted-mean point across the subregions.
| # | feature | importance | heterogeneity | #regions | regional analysis |
|---|---|---|---|---|---|
| 1 | x_3 | 0.6500 | 0.0000 | 1 | below the search threshold → |
| 2 | x_0 | 0.2886 | 0.0000 | 7 | split into 4 regions → |
| 3 | x_2 | 0.2817 | 0.2993 | 7 | split found — rejected by the decision sequence → |
| 4 | x_1 | 0.0015 | 0.0808 | · | not plotted (below the coverage cut) |
The plotted features carry 100% of the total importance mass (target 80%, ceiling top_k = 5).
The decision sequence. Starting from the global curves, each round applies the split with the largest explained-variance gain, measured on top of the splits above it, and stops when no remaining split adds at least 1.0%. A real split (its heterogeneity does drop) can still add nothing — or even hurt, by double-counting — when its variance is already explained by an earlier split.
| step | regions | heterogeneity | explained variance |
|---|---|---|---|
| global effects (GAM) | · | · | 84.9% |
| + split x_0 (on x_1, x_2) | 4 | 0.293 → 0.000 | +14.9% → 99.8% |
| rejected · x_2 (on x_0) | 4 | 0.299 → 0.165 | -9.8% — redundant (variance already explained) |
Bar view — importance and heterogeneity
2 · Regional analysis — the final CALM
The selected snapshot: global effects everywhere except the accepted splits. Features in descending importance — a split feature enters as one group at the instance-weighted mean of its subregions. The split features' global counterparts are in the baseline section at the end.
2.1 · x_3
Global effect
Regional effects
Heterogeneity below the threshold — the mean effect tells the whole story; find_regions was skipped.
2.2 · x_0
Split on x_1, x_2 into 4 regions — worth +14.9% of explained variance on top of the splits above it; importance and heterogeneity here are the instance-weighted means over the subregions. The global counterpart is in the baseline.
Partition tree
Feature 0 - Full partition tree:
🌳 Full Tree Structure:
───────────────────────
x_0 🔹 [id: 0 | heter: 0.29 | inst: 1000 | w: 1.00]
x_2 < -0.00 🔹 [id: 1 | heter: 0.05 | inst: 494 | w: 0.49]
x_1 < -0.00 🔹 [id: 2 | heter: 0.00 | inst: 243 | w: 0.24]
x_1 ≥ -0.00 🔹 [id: 3 | heter: 0.00 | inst: 251 | w: 0.25]
x_2 ≥ -0.00 🔹 [id: 4 | heter: 0.05 | inst: 506 | w: 0.51]
x_1 < -0.00 🔹 [id: 5 | heter: 0.00 | inst: 268 | w: 0.27]
x_1 ≥ -0.00 🔹 [id: 6 | heter: 0.00 | inst: 238 | w: 0.24]
--------------------------------------------------
Feature 0 - Statistics per tree level:
🌳 Tree Summary:
─────────────────
Level 0🔹heter: 0.29
Level 1🔹heter: 0.05 | 🔻0.25 (84.50%)
Level 2🔹heter: 0.00 | 🔻0.05 (100.00%)
2.3 · x_2
Global effect
Regional effects
A split on x_0 into 4 regions was found (heterogeneity 0.299 → 0.165), but the decision sequence skips it: it adds no explained variance beyond the splits kept there — the same variance is already read elsewhere. The regional plots are omitted; reproduce them with find_regions.
3 · Global baseline — without regions
What you would believe about the split features without the regional analysis: their global mean effects, with the heterogeneity the accepted splits just explained still hiding inside the band. Compare with their subregions in the regional analysis above.
x_0
