Computes V = Sigma_post %*% G %*% Sigma_post, the mixed-model form of the
YWH correction (see ywh_adjust()). The ALD working-likelihood score for
observation i is s_i = (1/sigma) x_i (tau - 1{resid_i < 0}), so the meat is
G = (1/sigma^2) sum_i x_i x_i' psi_i^2 (independence) or
G = (1/sigma^2) sum_g (sum_{i in g} x_i psi_i)(...)' (cluster-robust),
with psi_i = tau - 1{resid_i < 0}.
Arguments
- Sigma_post
Posterior covariance of the fixed effects (K x K).
- X
Fixed-effect design matrix (N x K).
- resid
Conditional residuals
y - (X beta + Z b)at the posterior median fit.- sigma
Posterior ALD scale (a positive scalar).
- tau
Quantile level.
- groups
Optional integer cluster index (length N).
NULLgives the independence meat.