ar_ensemble: AR Ensemble
In JuGross/ensAR: Autoregressive postprocessing methods for ensemble forecasts

Description Usage Arguments Value Note Author(s) Examples

For each member of a forecast ensemble, a corresponding autoregressive modified member is generated.

1	ar_ensemble(ens, obs_col, mem_col, train = 90, skip = 0)

`ens`	A data frame with one observation column, at least one forecast column, and at least one additional column (e.g. date).
`obs_col`	The observation column.
`mem_col`	The column(s) of the forecast members(s).
`train`	The length of the rolling training period used for fitting an autoregressive process.
`skip`	A number corresponding to the forecast ahead time (0 for ahead times not greater than 24 hours, 1 for ahead times greater than 24 hours and not greater than 48 hours, and so on).

A list with four elements:

observation: A numeric vector containing the observations for the forecast period (original period except for the first train dates.
forecast: A data frame containing the AR modified forecasts.
variance: A data frame containing variance estimates corresponding to the forecasts.
additional: A data frame containg the variables (columns) of ens which were not specified by obs_col and mem_col.

It is assumed that in each row of the data frame ens the forecast matches the observation, i.e. for each row the difference between forecast and observation is the forecast error.

J. Gross, A. Moeller.

ar_ensemble(ens = Magdeburg[1:(90 + 1), -c(57,58)],
    obs_col = 6, mem_col = 7:56)
ar_ensemble(ens = Magdeburg48[1:(90 + 1), -c(57,58)],
    obs_col = 6, mem_col = 7:56, skip = 1)