(Re-)fit an already-constructed MLPKriging object on new data. The MLP architecture and kernel are kept from construction.
# S3 method for MLPKriging
fit(
object,
y,
X,
regmodel = "constant",
normalize = FALSE,
optim = "BFGS+Adam",
objective = "LL",
parameters = NULL,
...
)No return value. MLPKriging object argument is modified.
MLPKriging object
numeric vector of observations (n)
numeric matrix of inputs (n x d)
trend: "constant", "linear", "quadratic"
logical; normalise inputs?
optimiser
"LL" (log-likelihood)
optional named list of tuning parameters
ignored