smooth (version 4.3.1)

rmultistep: Multiple steps ahead forecast errors

Description

The function extracts 1 to h steps ahead forecast errors from the model.

Usage

rmultistep(object, h = 10, error = c("default", "additive",
  "multiplicative"), ...)

Value

The matrix with observations in rows and h steps ahead values in columns. So, the first row corresponds to the forecast produced from the 0th observation from 1 to h steps ahead.

Arguments

object

Model estimated using one of the forecasting functions.

h

The forecasting horizon to use.

error

Defines what type of error to return. "default" means returning the one used in the original model. "additive" is to return e_t = y_t - mu_t. Finally, "multiplicative" will return e_t = (y_t - mu_t) / mu_t.

...

Currently nothing is accepted via ellipsis.

Author

Ivan Svetunkov,

Details

The errors correspond to the error term epsilon_t in the ETS models. Don't forget that different models make different assumptions about epsilon_t and / or 1+epsilon_t.

See Also

residuals,

Examples

x <- rnorm(100,0,1)
ourModel <- adam(x)
rmultistep(ourModel, h=13)