/savitzky_golay_with_errors

Improving the Savitzky-Golay filtering in the presence of noisy data

Primary LanguagePython

Savitzky Golay filter for data with errorbars

A Savitzky–Golay filter (Savitzky, A.; Golay, M.J.E., 1964, "Smoothing and Differentiation of Data by Simplified Least Squares Procedures") is often applied to data for the purpose of smoothing the data without greatly distorting the signal.

This kind of filter has been often used in the scientific literature. However almost all data inherently comes with noise, and the noise properties can differ from point to point. This python script improves upon the traditional Savitzky-Golay filter by accounting for errors or covariance in the data.

The inputs and arguments are all modelled after scipy.signal.savgol_filter