I've published some R
code on CRAN
that can be used to calculate the incredible higher-order LQ
decomposition (incredible HOLQ), the incredible singular value
decomposition (ISVD), and the incredible higher-order polar
decomposition (IHOP). The HOLQ can be used to (1) derive the maximum
likelihood estimates of the covariance matrices under the array normal
model, (2) run a likelihood ratio test in separable covariance models,
and (3) calculate AIC's and BIC's for separable covariance
models. There are also a few useful array and matrix functions. A
description of the methods can be found in:
Gerard, D. , & Hoff, P. (2016). A higher-order LQ decomposition for separable covariance models. Linear Algebra and its Applications, 505, 57-84. [Link to LAA] [Link to arXiv] [bib]
I also provide a brief vignette on how to use the functions available in the R
code. To download, simply type in R
:
install.packages("tensr")
Alternatively, you can download it directly from CRAN.
Please provide us with questions or comments: David Gerard (dcgerard) or Peter Hoff (pdhoff) -- @uchicago.edu and @uw.edu, respectively.