Compute each event's total random-effect contribution from a flattened re_effect[D, total_re_levels] matrix (as emitted by the generic Stan models' generated re_effect, or reconstructed R-side from posterior draws of z_re/tau_re/L_corr_re). This is the SINGLE generic helper every downstream mu-reconstruction site should call instead of hand-summing hospital_effect/patient_effect/admission_effect.

Compute each event's total random-effect contribution from a flattened re_effect[D, total_re_levels] matrix (as emitted by the generic Stan models' generated re_effect, or reconstructed R-side from posterior draws of z_re/tau_re/L_corr_re). This is the SINGLE generic helper every downstream mu-reconstruction site should call instead of hand-summing hospital_effect/patient_effect/admission_effect.

Usage

re_contribution(re_effect, flat_group_index)

Arguments

re_effect

D x total_re_levels numeric matrix (one posterior draw).

flat_group_index

n_events x R integer matrix (see prepare_random_effects()$flat_group_index), or a single event's R-length integer vector.

Value

If flat_group_index is a matrix: n_events x D matrix of summed RE contributions. If a vector: length-D vector for one event.