Predict using vector autoregressive spatio-temporal model

Predicts values given new covariates using a tinyVAST model

Usage

# S3 method for tinyVAST
predict(
  object,
  newdata,
  remove_origdata = FALSE,
  what = c("mu_g", "p_g", "palpha_g", "pgamma_g", "pepsilon_g", "pomega_g", "pxi_g"),
  se.fit = FALSE,
  ...
)

Arguments

object

Output from tinyVAST().

newdata

New data-frame of independent variables used to predict the response.

remove_origdata

Whether to eliminate the original data from the TMB object, thereby speeding up the TMB object construction. However, this also eliminates information about random-effect variance, and is not appropriate when requesting predictive standard errors or epsilon bias-correction.

what

What REPORTed object to output, where p_g is the linear predictor, mu_g is the inverse-linked transformed predictor. and others are additive components of the linear predictor.

se.fit

Calculate standard errors?

...

Not used.