ADPF: Use Least Squares Polynomial Regression and Statistical Testing to Improve Savitzky-Golay

This function takes a vector or matrix of data and smooths the data with an improved Savitzky Golay transform. The Savitzky-Golay method for data smoothing and differentiation calculates convolution weights using Gram polynomials that exactly reproduce the results of least-squares polynomial regression. Use of the Savitzky-Golay method requires specification of both filter length and polynomial degree to calculate convolution weights. For maximum smoothing of statistical noise in data, polynomials with low degrees are desirable, while a high polynomial degree is necessary for accurate reproduction of peaks in the data. Extension of the least-squares regression formalism with statistical testing of additional terms of polynomial degree to a heuristically chosen minimum for each data window leads to an adaptive-degree polynomial filter (ADPF). Based on noise reduction for data that consist of pure noise and on signal reproduction for data that is purely signal, ADPF performed nearly as well as the optimally chosen fixed-degree Savitzky-Golay filter and outperformed sub-optimally chosen Savitzky-Golay filters. For synthetic data consisting of noise and signal, ADPF outperformed both optimally chosen and sub-optimally chosen fixed-degree Savitzky-Golay filters. See Barak, P. (1995) <doi:10.1021/ac00113a006> for more information.

Getting started

Package details

AuthorPhillip Barak [aut], Samuel Kruse [cre, aut]
MaintainerSamuel Kruse <[email protected]>
LicenseGPL-3
Version0.0.1
Package repositoryView on CRAN
Installation Install the latest version of this package by entering the following in R:
install.packages("ADPF")

Try the ADPF package in your browser

Any scripts or data that you put into this service are public.

ADPF documentation built on Sept. 14, 2017, 1:02 a.m.