robfilter: Robust Time Series Filters

Implementations for several robust procedures that allow for (online) extraction of the signal of univariate or multivariate time series by applying robust regression techniques to a moving time window are provided. Included are univariate filtering procedures based on repeated-median regression as well as hybrid and trimmed filters derived from it; see Schettlinger et al. (2006) <doi:10.1515/BMT.2006.010>. The adaptive online repeated median by Schettlinger et al. (2010) <doi:10.1002/acs.1105> and the slope comparing adaptive repeated median by Borowski and Fried (2013) <doi:10.1007/s11222-013-9391-7> choose the width of the moving time window adaptively. Multivariate versions are also provided; see Borowski et al. (2009) <doi:10.1080/03610910802514972> for a multivariate online adaptive repeated median and Borowski (2012) <doi:10.17877/DE290R-14393> for a multivariate slope comparing adaptive repeated median. Furthermore, a repeated-median based filter with automatic outlier replacement and shift detection is provided; see Fried (2004) <doi:10.1080/10485250410001656444>.

Package details

AuthorRoland Fried [aut, cre], Karen Schettlinger [aut], Matthias Borowski [aut], Robin Nunkesser [ctb], Thorsten Bernholt [ctb]
MaintainerRoland Fried <>
LicenseGPL (>= 2)
Package repositoryView on CRAN
Installation Install the latest version of this package by entering the following in R:

Try the robfilter package in your browser

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

robfilter documentation built on Nov. 21, 2019, 5:07 p.m.