This convenience wrapper assembles the ingredients the Gaussian process (GP)
model needs so downstream functions can work with a single, tidy bundle.
Usage
gp_build_state(obs_data, coordinates, hyperparameters, n, nt, period = 52)
Arguments
- obs_data
Observation data
- coordinates
Site coordinates
- hyperparameters
Vector of hyperparameters
- n
N sites
- nt
N times
- period
Period
Value
List of pre-computaed GP elements
Details
It builds the spatial and temporal kernels, indexes the observed counts,
prepares working responses and variances for the Poisson log-scale
approximation, and returns the matrices and closures required for GP
inference.