quantwave_core/features/
regime_probs.rs

1//! Regime probability features (soft labels) for ML pipelines.
2//!
3//! Provides a simple way to turn hard regime labels or basic clustering into
4//! probability-like vectors that are excellent stationary features.
5//!
6//! This is a thin wrapper; richer soft outputs (HMM forward probs, GMM responsibilities)
7//! can be wired here later when the regimes module exposes them cleanly.
8//!
9//! Source: regimes/mod.rs (MarketRegime) + quantwave-4ub research notes (regime probs as meta-features).
10
11use crate::regimes::MarketRegime;
12
13#[derive(Debug, Clone, Copy, PartialEq)]
14pub struct RegimeProbFeatures {
15    /// 5-dimensional soft vector (Bull, Bear, Crisis, Steady, Other/Cluster)
16    pub probs: [f64; 5],
17    pub hard_label: MarketRegime,
18}
19
20pub fn regime_to_prob_features(regime: MarketRegime) -> RegimeProbFeatures {
21    let mut probs = [0.05; 5]; // small uniform prior
22    let idx = match regime {
23        MarketRegime::Bull => 0,
24        MarketRegime::Bear => 1,
25        MarketRegime::Crisis => 2,
26        MarketRegime::Steady => 3,
27        MarketRegime::Cluster(c) => 4.min(c as usize),
28    };
29    probs[idx] = 0.80;
30    // normalize roughly
31    let sum: f64 = probs.iter().sum();
32    for p in &mut probs {
33        *p /= sum;
34    }
35    RegimeProbFeatures { probs, hard_label: regime }
36}
37
38#[cfg(test)]
39mod tests {
40    use super::*;
41
42    #[test]
43    fn test_regime_prob_features_bull() {
44        let f = regime_to_prob_features(MarketRegime::Bull);
45        assert!(f.probs[0] > 0.7);
46        assert_eq!(f.hard_label, MarketRegime::Bull);
47    }
48}
quantwave_core/features/regime_probs.rs

quantwave_core/features/
regime_probs.rs
