ddpca: Diagonally Dominant Principal Component Analysis

Consider the problem of decomposing a large covariance matrix into a low rank matrix plus a diagonally dominant matrix. This problem is called Diagonally Dominant Principal Component Analysis (DD-PCA) in the reference Ke, Z., Xue, L. and Yang, F. (2019) <arXiv:1906.00051>. DD-PCA can be used in covariance matrix estimation and global detection in multiple testing. This package implements DD-PCA using both convex approach and non-convex approach; Convex approach refers to solving a convex relaxation of the original problem using Alternating Direction Method of Multipliers (ADMM), while non-convex approach resorts to an iterative projection algorithm. This package also implements two global testing methods proposed in the reference.

Version: 1.0
Imports: RSpectra, Matrix, quantreg, MASS
Published: 2019-07-19
Author: Zheng Tracy Ke Lingzhou Xue Fan Yang
Maintainer: Fan Yang <fyang1 at uchicago.edu>
License: GPL-2
NeedsCompilation: no
CRAN checks: ddpca results


Reference manual: ddpca.pdf
Package source: ddpca_1.0.tar.gz
Windows binaries: r-devel: ddpca_1.0.zip, r-release: ddpca_1.0.zip, r-oldrel: ddpca_1.0.zip
OS X binaries: r-release: ddpca_1.0.tgz, r-oldrel: ddpca_1.0.tgz


Please use the canonical form https://CRAN.R-project.org/package=ddpca to link to this page.