Extract and tabulate parameter estimates and metadata from a fitted SPoRC model

Joins the parameter list, factor map, and sdreport to produce two tidy data frames: one for estimated parameters (with initial values, posterior estimates, standard errors, and absolute gradients) and one for fixed (non-estimated) parameters. Parameter indices are re-sequenced to match the sequential numbering used internally by RTMB's sdreport, and map entries of NA (i.e., parameters fixed via mapping) are labelled "NE" (not estimated).

Usage

get_par_est_info(parameters, mapping, sd_rep)

Arguments

parameters

Named list of parameter starting values passed to RTMB::MakeADFun.

mapping

Named list of factor maps passed to RTMB::MakeADFun. Parameters absent from mapping are treated as freely estimated. Elements with NA factor levels are treated as fixed.

sd_rep

sdreport object returned by RTMB::sdreport. Must contain $par.fixed, $par.random, $cov.fixed, $diag.cov.random, and $gradient.fixed.

Value

A named list with two elements:

est_pars

Data frame of estimated parameters, with columns Par (parameter name), Est_Val (estimated value on the native scale), SE_Val (standard error), Abs_Grad_Val (absolute gradient; NA for random effects), Init_Val (starting value), and dimension columns Dim1, Dim2, ... indicating the array indices of each element.

non_est_pars

Data frame of non-estimated (fixed) parameters with the same dimension and Init_Val columns, plus map = "NE" indicating they were excluded from estimation.

See also

Other Utility: fit_model(), get_logistN_Sigma(), post_optim_sanity_checks(), rho_trans(), set_data_indicator_unused()