Expand description
Regime probability features (soft labels) for ML pipelines.
Provides a simple way to turn hard regime labels or basic clustering into probability-like vectors that are excellent stationary features.
This is a thin wrapper; richer soft outputs (HMM forward probs, GMM responsibilities) can be wired here later when the regimes module exposes them cleanly.
Source: regimes/mod.rs (MarketRegime) + quantwave-4ub research notes (regime probs as meta-features).