About

This repository is an R package implementing methods from the following two publications:

• Emily C. Hector and Peter X.-K. Song (2020a). A distributed and integrated method of moments for high-dimensional correlated data analysis. Journal of the American Statistical Association, pages 1–14. doi: 10.1080/01621459.2020.1736082.
• Emily C. Hector and Peter X.-K. Song (2020b). Doubly distributed supervised learning and inference with high-dimensional correlated outcomes. Journal of Machine Learning Research, 21:1–35.

Briefly, the method performs regression analysis of high-dimensional correlated responses. It divides data into blocks according to supplied indicators, analyses blocks using composite likelihood or generalized estimating equations (GEE), and combines blocks using a meta-estimator asymptotically equivalent to the optimal generalized method of moments (GMM) equation.

Please email ehector@ncsu.edu with any questions or bug-reports.

Installation

The DDIMM R package can be installed in one of two ways:

• from the downloaded gzipped tarball as R CMD INSTALL DDIMM_1.0-1.tar.gz
• from the downloaded and renamed DDIMM folder as R CMD build DDIMM and R CMD INSTALL DDIMM_1.0-1.tar.gz

Please make sure to have all packages listed in the DESCRIPTION file already installed. If you encounter a library not found error for lgfortran, please try installing gfortran from here: https://cran.r-project.org/bin/macosx/tools/.

Citation

If you use the DDIMM R package, please consider citing the relevant manuscript: Hector & Song (2020a) or Hector & Song (2020b).

References

The posdef.matrix function was written by Ravi Varadhan: https://stat.ethz.ch/pipermail/r-help/2008-February/153708.

The GEE implementation is through the R package geepack: Højsgaard, S., Halekoh, U., Yan, J. (2006). The R package geepack for generalized estimating equations. Journal of Statistical Software, 15(2):1–11.