pastas.noisemodels.ArNoiseModel.get_correction#

pastas.noisemodels.ArNoiseModel.get_correction(res: pandas.Series, p: pastas.typing.ArrayLike, tindex: pandas.DatetimeIndex) → pandas.Series#

Get correction for a forecast using the noise model.

Parameters:

  • res (Series) – The residual series.

  • p (ArrayLike) – The parameters of the noise model.

  • tindex (DatetimeIndex) – The index of the forecast.

Returns:

The correction to the forecast.

Return type:

Series

Notes

The correction is calculated as:

\[\begin{split}correction = \\exp(-\\Delta t / \\alpha) * last_residual\end{split}\]

where \(\\Delta t\) is the time difference between the last observation and the forecast, and \(\\alpha\) is the noise parameter.