Skip to contents

Get Plot of Negative Log Likelihood Values

get_nLL_plot.Rd

Extracts, weights, and visualises all negative log-likelihood (nLL) components from one or more SPoRC model runs. Likelihood weights stored in the data list are applied to the relevant components (catch, indices, recruitment, tagging, fishing mortality) before plotting, so reported values reflect the actual contribution of each component to the joint nLL. Components with a value of zero are silently excluded from both the plot and table.

Usage

get_nLL_plot(data, rep, model_names)

Arguments

data

List of length n_models, where each element is a SPoRC data list. Likelihood weights are extracted from Wt_Catch, Wt_F, Wt_Rec, Wt_Tagging, Wt_SrvIdx, and Wt_FishIdx.

rep

List of length n_models, where each element is a SPoRC report list (i.e. the output of obj$report() after optimisation). The following nLL components are extracted: jnLL, h_nLL, M_nLL, rec_region_prop_nLL, Rec_nLL, Init_Rec_nLL, sel_nLL, conv_fish_tag_nLL, Catch_nLL, Fmort_nLL, srv_q_nLL, fish_q_nLL, SrvIdx_nLL, FishIdx_nLL, TagRep_nLL, Movement_nLL, SrvAgeComps_nLL, FishAgeComps_nLL, SrvLenComps_nLL, FishLenComps_nLL, and the population-specific counterparts Catch_pop_nLL, FishIdx_pop_nLL, SrvIdx_pop_nLL, FishAgeComps_pop_nLL, FishLenComps_pop_nLL, SrvAgeComps_pop_nLL, SrvLenComps_pop_nLL. Missing components are handled via safe_extract.

model_names

Character vector of length n_models giving display names for each model run. Used as facet labels on the bar chart.

Value

A list of two objects:

[[1]] nLL_plot

A stacked bar chart (ggplot) with one facet per model. Bars show the weighted nLL contribution of each component, coloured by component type (Prior, Penalty, Catch, Index, Age, Length, Tagging, jnLL). Components with value 0 are excluded.

[[2]] table_plot

A ggdraw table (via gridExtra and cowplot) showing weighted nLL values in wide format, with one column per non-zero component and one row per model.

See also

Other Model Diagnostics: do_jitter(), do_likelihood_profile(), do_retrospective(), do_runs_test(), get_catch_fits_plot(), get_comp_prop(), get_idx_fits(), get_idx_fits_plot(), get_model_rep_from_mcmc(), get_osa(), get_retrospective_plot(), get_retrospective_relative_difference(), plot_resids()

Examples

if (FALSE) { # \dontrun{
  out <- get_nLL_plot(
    data        = list(data1, data2),
    rep         = list(rep1, rep2),
    model_names = c("Base", "Sensitivity")
  )
  out[[1]]  # bar chart
  out[[2]]  # table
} # }