Fit hyperparmeters

Usage

fit(
  obs_data,
  coordinates,
  nt,
  period,
  n_sites,
  mask_prop = 0.2,
  verbose = FALSE,
  par0 = c(1, 5, 100),
  lower = c(1e-04, 0.8, 52 * 1.5),
  upper = c(2, 10, 500)
)

Arguments

obs_data: Data frame of observations with at least `id`, `y_obs`, `mu_infer`, and `f_infer` columns.
coordinates: Data frame of site coordinates containing an `id` column that matches the site identifiers in `obs_data`.
nt: Number of time points per site.
period: Period used for the temporal kernel.
n_sites: Number of sites to sample for optimisation.
mask_prop: Proportion of observed points per site to hold out.
verbose: Logical; whether to print objective values during optimisation.
par0: Initial hyperparameter values passed to `optim()`.
lower: Lower bounds for the hyperparameters passed to `optim()`.
upper: Upper bounds for the hyperparameters passed to `optim()`.

Value

A list as returned by `stats::optim()`.