Extend a series using year over year growth

Usage

yoy_to_lev(yoy_gr, hist_lev, smooth_span = 0)

Arguments

yoy_gr: ts-boxable object containing year over year growth rates
hist_lev: ts-boxable object containing the history in levels for forecast and at least one year of history (in percent)
smooth_span: extent of smoothing between 0-1 (default: 0, no smoothing). Effect depends on the length of the time series. Low smooth_span value may trigger warning, if time series too short.

Value

object of the same type as hist_lev extended with year over year growth

Details

This function only works for univariate time series and requires that the growth rates are available for at least the last year of history. Year-over-year growth rates propagate the fluctuations of the base period into the extension period. This can be mitigated by smoothing the extension. The inputs should not contain missing values.

Examples

gr <- quarterly_data_example |>
  tsbox::ts_long() |>
  dplyr::filter(id == "E_NF_HI") |>
  tsbox::ts_pcy()
lev <- quarterly_data_example |>
  tsbox::ts_long() |>
  dplyr::filter(id == "ECT_HI")
res1 <- yoy_to_lev(gr, lev |> dplyr::filter(time <= "2010-01-01"))
res2 <- yoy_to_lev(gr, lev |> dplyr::filter(time <= "2010-01-01"), 1/8)
tsbox::ts_plot(lev, res1, res2)