Smoothing with Savitzky–Golay filter
I suggest implementing the "Savitzky–Golay" filter for time series smoothing purposes.
https://en.wikipedia.org/wiki/Savitzky%E2%80%93Golay_filter
I suggest implementing the "Savitzky–Golay" filter for time series smoothing purposes.
https://en.wikipedia.org/wiki/Savitzky%E2%80%93Golay_filter
Comentários
The Savitzky–Golay filter is now available in NxLOCREG(.) function.
https://support.numxl.com/hc/en-us/articles/360031283631-NxLOCREG-Local-or-moving-polynomial-regression
Outside econometric, LOESS is known and commonly referred to as Savitzky–Golay filter. Savitzky–Golay filter was proposed 15 years before LOESS.
NumXL 1.66 PARSON supports both kernel and localized regression, so extending it to support time series smoothing is next on the release-plan.
iniciar sessão para comentar.