Computes the noise series (output minus fitted signal) from a transfer function model.
# S3 method for ssm
noise(mdl, diff = TRUE, exp = FALSE, ...)noise(mdl, ...)
# S3 method for tfm
noise(mdl, y = NULL, diff = TRUE, exp = FALSE, envir = parent.frame(), ...)
A ts object containing the noise series, computed as
output minus all transfer function and regressor effects.
A tfm object.
Logical. If TRUE (default), returns differenced noise series (stationary). If FALSE, returns noise in original scale.
Logical. If TRUE, applies exponential transformation (inverse
of log). Only relevant when diff = FALSE and Box-Cox transformation
was used (bc = TRUE).
Additional arguments.
Optional ts object for alternative output series.
Environment for evaluation. NULL uses calling environment.
The noise represents the component of the output not explained by the
transfer functions and exogenous regressors. When diff = TRUE,
the differencing operator from the noise model is applied, resulting in
a stationary series suitable for ARMA modeling.
signal.tfm, residuals.tfm, tfm