Pseudo out-of-sample Forecasting based on Expanding Window

Usage

# S3 method for class 'marbayes'
forecast_expand(
  object,
  n_ahead,
  y_test,
  level = 0.05,
  newxreg,
  num_thread = 1,
  med = FALSE,
  lpl = FALSE,
  mcmc = TRUE,
  use_fit = TRUE,
  verbose = FALSE,
  ...
)

Arguments

object: Model object
n_ahead: Step to forecast in rolling window scheme
y_test: Test data to be compared.
level: Specify alpha of confidence interval level 100(1 - alpha) percentage. By default, .05.
newxreg: New values for exogenous variables.
num_thread: Number of threads
med: If TRUE, use median of forecast draws instead of mean (default).
lpl: Compute log-predictive likelihood (LPL). By default, FALSE.
mcmc: If TRUE, run new MCMC in new windows. By default, TRUE.
use_fit: Use object result for the first window. By default, TRUE.
verbose: Print the progress bar in the console. By default, FALSE.
...: Additional arguments