EbayesThresh:
R and S-PLUS software for Empirical Bayes thresholding.
by Iain M. Johnstone and Bernard W.
Silverman, August 2002, last revised
This software complements
the papers "Needles and straw in haystacks: Empirical Bayes approaches to
thresholding a possibly sparse sequence", "Empirical Bayes selection
of wavelet thresholds", and "EbayesThresh: R software for Empirical
Bayes thresholding", by Iain M. Johnstone and Bernard W. Silverman. All the papers are available here. The last
of these papers gives a
general description of the software and some details both of the
general methodology and of some specific technical matters.
The R package (current
version 1.3.0) and documentation is available as the contributed package EbayesThresh from the Comprehensive R Archive
Network (CRAN). It can also be downloaded directly from
within R.
The master routine in the
software for processing single sequences of data is the function ebayesthresh. The routine ebayesthresh.wavelet
applies the approach to
wavelet transforms obtained within the S+WAVELETS or R waveslim or wavethresh
modules. If wavelet transforms are obtained using other software, the routine
will not be applicable directly, but should still provide a model for the user
to write their own wavelet smoothing routine making use of the function ebayesthresh.
The software may be
downloaded and used freely for academic purposes, provided its use is
acknowledged. Commercial use is not allowed without the permission of the author. Please bring any problems
or errors to the author's attention.
Anestis Antoniadis and
Maarten Jansen have provided a MATLAB version of the package, available here.