Weight Composition Likelihood

Computes likelihood for observed weight compositions using the PLA matrix and a precomputed rebinning matrix. The prediction pipeline is: catch_at_age -> PLA -> pred_at_length -> rebin_matrix -> pred_at_weight. Gradients propagate through PLA to growth parameters when estimated.

Usage

get_weight_like(
  wf_obs_flat,
  wf_obs_ints,
  wf_obs_prop,
  catch_pred_fya,
  pla,
  wf_rebin_matrix,
  wf_n_f,
  wf_fishery_f,
  wf_year_fi,
  wf_n_fi,
  wf_minbin,
  wf_maxbin,
  removal_switch_f,
  wf_switch,
  n_wt,
  n_wf,
  log_wf_tau
)

Arguments

wf_obs_flat

numeric vector of unrounded weight comp counts (wf_switch=1).

wf_obs_ints

integer vector of integer weight comp counts (wf_switch=3).

wf_obs_prop

numeric vector of weight comp proportions (wf_switch=2).

catch_pred_fya

3D array [n_fishery, n_year, n_age] of predicted catch-at-age from do_dynamics().

pla

matrix [n_len, n_age] probability-of-length-at-age from get_pla(). On the AD tape when growth parameters are estimated.

wf_rebin_matrix

matrix [n_wt, n_len] precomputed rebinning weights from prep_wf_data().

wf_n_f

integer vector [n_fishery] of observation counts per fishery.

wf_fishery_f

integer vector of fishery indices with WF data.

wf_year_fi

list of integer vectors of year indices per fishery.

wf_n_fi

list of integer vectors of sample sizes per fishery.

wf_minbin

integer vector [n_fishery] minimum weight bin index.

wf_maxbin

integer vector [n_fishery] maximum weight bin index.

removal_switch_f

integer vector [n_fishery] removal flags.

wf_switch

integer likelihood type (1=multinomial, 2=Dirichlet, 3=DM).

n_wt

integer number of weight bins.

n_wf

integer total number of WF observations.

log_wf_tau

numeric vector [n_fishery] log-scale variance adjustment.

Value

numeric vector of negative log-likelihood contributions, one per observation.