Struct GibbsSampler 

pub struct GibbsSampler { /* private fields */ }

Gibbs sampler for Student’s t prior inference.

Implementations

impl GibbsSampler

pub fn new(l_r: &Matrix9, sigma: f64, seed: u64) -> Self

Create a new Gibbs sampler with precomputed factorizations.

Arguments

• l_r - Cholesky factor of R (correlation matrix)
• sigma - Calibrated prior scale
• seed - Deterministic RNG seed

pub fn run( &mut self, delta_obs: &Vector9, sigma_n: &Matrix9, theta: f64, ) -> GibbsResult

Run Gibbs sampler and return posterior summaries.

Arguments

• delta_obs - Observed quantile differences Δ
• sigma_n - Likelihood covariance Σ_n
• theta - Effect threshold for leak probability

v5.6: Extended to sample (δ, λ, κ) for robust t-likelihood.