This method is primarily intended to be used only for models fit using optimization. For models fit using MCMC or one of the variational approximations, see posterior_predict.

# S3 method for stanreg
predict(
  object,
  ...,
  newdata = NULL,
  type = c("link", "response"),
  se.fit = FALSE
)

Arguments

object

A fitted model object returned by one of the rstanarm modeling functions. See stanreg-objects.

...

Ignored.

newdata

Optionally, a data frame in which to look for variables with which to predict. If omitted, the model matrix is used.

type

The type of prediction. The default 'link' is on the scale of the linear predictors; the alternative 'response' is on the scale of the response variable.

se.fit

A logical scalar indicating if standard errors should be returned. The default is FALSE.

Value

A vector if se.fit is FALSE and a list if se.fit is TRUE.

See also