Forms a cheap estimate of the latent log-intensity field as the per-site
centred (and optionally scaled) log1p of the observed counts. Missing cells
are filled with the per-site mean (zero after centring) – a neutral fill
that carries no signal of its own. See refine = TRUE in
infer_kernel_params() for a correlation-aware fill.
Usage
build_plugin_field(obs_data, n, nt, value = "y_obs", standardise = TRUE)
Arguments
- obs_data
Data frame with id (site), t (time) and the count column
named by value.
- n
Number of sites.
- nt
Number of time points.
- value
Name of the count column (default "y_obs").
- standardise
Logical; scale each site to unit variance after centring
(default TRUE).
Value
A numeric vector of length n * nt, ordered sites x times (time
fastest).
Details
Per-site centring removes the site intercept mu_s; per-site scaling
homogenises per-site variances so a single global sigma^2 and the
correlation kernels apply.