Quantitative Signals & Scoring

Regime Detection Probability

The statistical likelihood that a quantitative signal correctly identifies a shift in market regime or insider trading behavior pattern, measured as the conditional probability that observed signal characteristics align with known regime markers.

Formula

RDP_{t} = P (Regime_{k} ∣ S_{t}) = \frac{P ( S _{t} ∣ Regime _{k} ) \cdot P ( Regime _{k} )}{\sum _{j = 1}^{K} P ( S _{t} ∣ Regime _{j} ) \cdot P ( Regime _{j} )}

\textRDP_t: Regime detection probability at time t
\textRegime_k: Discrete regime class k from a predefined set of K regimes
\mathbfS_t: Vector of observed signal features at time t
P(\mathbfS_t | \textRegime_k): Likelihood of observing signal features given regime k
P(\textRegime_k): Prior probability of regime k based on historical frequency

Related terms

See it in the data