Select balanced rows from a combined grid sample

Selects unique posterior rows with counts that differ by at most one across grid cells. The selected rows are returned in a reproducibly shuffled order.

Usage

select_balanced_grid_iters(metadata, n_draws, seed, expected_cells = NULL)

Arguments

metadata

Data frame with unique integer draw identifiers and a cell identifier for every available row.

n_draws

Positive integer number of rows to select.

seed

Integer random seed.

expected_cells

Optional vector giving the complete expected set and order of grid cells. By default, sorted unique cells in metadata are used.

Value

An integer vector of values from metadata$draw.

See also

Other grid workflows: grid_mcmc_to_tmbfit(), grid_to_tmbfit(), make_mle_grid_parameters(), run_grid(), run_grid_mcmc(), sample_grid(), summarise_mle_grid(), validate_mle_grid_state_records()