Top pick: Sliced Wasserstein dependency (POT (MIT)). Fills the gap: Stability across regimes — which the production baseline (Pearson + Spearman + h_norm + Tier-2 MI-vs-target) cannot detect.
This category is Tier B (adjacent measurement) — doesn't directly score orthogonality. Top P4 metric shown below for reference.
Metric: Sliced Wasserstein dependency ·
FOSS: POT (MIT) ·
Effort: Moderate ·
Gap closed: Regime stability (Tier B adjacent — Batch A skips)
→ See the full priority list — 30 metrics ranked P1 → P4 across 6 categories.
Lane = how usable the metric is today. Top lane = production-baseline duplicates (Pearson, Spearman). Below = permissive FOSS / GPL / unmaintained / code-released-but-not-packaged / no-FOSS. Per-circle text = first 3 letters of language (Pyt / R / Jav).
| Metric | Priority | Library | License | Lang | Maint | Complexity | Param-free? | URL |
|---|---|---|---|---|---|---|---|---|
| Robust BOCPD | P4 | Author repo (Altamirano-Briol-Knoblauch 2023) | varies | Python | ⚠️ | O(N) | yes | ↗ |
| Robust dCor (Leyder 2024) | — | Author code (not packaged) | varies | R | ⚠️ | O(N^2) | no | ↗ |
| CRQA | P4 | crqa / PyRQA | GPL / Apache-2 | R / Python | ✅ | O(N^2) | yes | ↗ |
| Sliced Wasserstein primitives | P4 | POT | MIT | Python | ✅ | O(N log N d L) | yes | ↗ |