PDP report
target y · 3 features · 2 plotted
1 · Overview — where to look
An additive surrogate read off the global curves reproduces 90.5% of the model's predicted variance; adding the regional plots kept by the decision sequence, 98.4%.
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 | level | 1.1028 | 0.3607 | 5 | split found — rejected by the decision sequence → |
| 2 | x1 | 0.2291 | 0.0757 | 5 | split into 3 regions → |
| 3 | x2 | 0.0069 | 0.2635 | · | not plotted (below the coverage cut) |
The plotted features carry 99% 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) | · | · | 90.5% |
| + split x1 (on level, x2) | 3 | 0.370 → 0.076 | +7.9% → 98.4% |
| rejected · level (on x1, x2) | 3 | 0.361 → 0.129 | -4.8% — 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 · level
Global effect
Regional effects
A split on x1, x2 into 3 regions was found (heterogeneity 0.361 → 0.129), 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.
2.2 · x1
Split on level, x2 into 3 regions — worth +7.9% 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 1 - Full partition tree:
🌳 Full Tree Structure:
───────────────────────
x1 🔹 [id: 0 | heter: 0.37 | inst: 1000 | w: 1.00]
x2 < 0.00 🔹 [id: 1 | heter: 0.00 | inst: 498 | w: 0.50]
x2 ≥ 0.00 🔹 [id: 2 | heter: 0.51 | inst: 502 | w: 0.50]
level = 0.00 🔹 [id: 3 | heter: 0.00 | inst: 233 | w: 0.23]
level ∈ {1.00, 2.00} 🔹 [id: 4 | heter: 0.28 | inst: 269 | w: 0.27]
--------------------------------------------------
Feature 1 - Statistics per tree level:
🌳 Tree Summary:
─────────────────
Level 0🔹heter: 0.37
Level 1🔹heter: 0.26 | 🔻0.11 (31.02%)
Level 2🔹heter: 0.08 | 🔻0.18 (70.34%)
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.
x1
