knitr::opts_chunk$set( collapse = TRUE, comment = "#>", fig.path = "README-" )
hose
is a package designed for working with higher-order spectral
estimators, which were first introduced in @gerard2017adaptive. These
estimators are based on the higher-order singular value decomposition
of @lathauwer2000multilinear and are useful when your data exhibit
tensor-specific structure, such as having approximately low
multilinear rank. This code will allow you to:
The main functions are:
get_c()
: Pre-format the data before applying mode-specific singular value
shrinkage.tensor_var_est()
: Estimate the variance of the data from multiple options.soft_coord()
: Estimate the underlying low-rank mean tensor via
soft-thresholding.If you find these methods useful, please cite
Gerard, David, and Peter Hoff. 2017. "Adaptive Higher-Order Spectral Estimators." Electron. J. Statist. 11 (2). The Institute of Mathematical Statistics; the Bernoulli Society: 3703--37. https://doi.org/10.1214/17-EJS1330.
Or, using BibTex:
@ARTICLE{gerard2017adaptive, AUTHOR = {David Gerard and Peter Hoff}, TITLE = {Adaptive higher-order spectral estimators}, JOURNAL = {Electron. J. Statist.}, FJOURNAL = {Electronic Journal of Statistics}, YEAR = {2017}, VOLUME = {11}, NUMBER = {2}, PAGES = {3703-3737}, ISSN = {1935-7524}, DOI = {10.1214/17-EJS1330}, SICI = {1935-7524(2017)11:2<3703:AHOSE>2.0.CO;2-Q}, }
You can install from CRAN in the usual way:
install.packages("hose")
Or, to install the latest (unstable) version, run the following code in R:
install.packages(c("tensr", "softImpute", "RMTstat", "devtools")) devtools::install_github("dcgerard/hose")
I've provided a vignette demonstrating the methods available in
hose
. You can find it here.
Or you can build the vignette on install with
install.packages("devtools") devtools::install_github("dcgerard/hose", build_vignettes = TRUE)
and access the vignette by running the following code in R:
utils::vignette("sure_example", package = "hose")
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.