yk748/cxreg: cxreg v1.1.4

cxreg v1.1.4 cxreg is an R package for complex-valued penalized regression and Gaussian graphical models. It provides efficient estimation and inference procedures for complex-valued lasso, graphical lasso, and high-dimensional spectral precision matrices. This release adds a comprehensive spectral inference vignette, makes the current release available on CRAN, and accompanies the publication of the cxreg software paper in the Journal of Open Source Software. New features Spectral inference vignette A new vignette, "Inference for Sparse Spectral Precision Matrices with cxreg," provides a complete worked example of the spectral inference pipeline introduced in the 1.1.x series. The vignette demonstrates the full workflow: select_m() — data-driven bandwidth selection via generalised cross-validation. fhat_at() — construction of the smoothed periodogram matrix at a selected Fourier frequency. cglasso() — estimation of a sparse spectral precision matrix. decglasso() — one-step debiasing of the CGLASSO estimator. var.cov() — asymptotic variance and pseudovariance estimation. spec.test() — entry-wise hypothesis testing and confidence regions. spec.fdr() — FDR-controlled multiple testing for spectral support recovery. The vignette complements the existing examples and provides an end-to-end illustration of inference for high-dimensional sparse spectral precision matrices. Available on CRAN cxreg v1.1.4 is now available from CRAN: https://CRAN.R-project.org/package=cxreg The current release can be installed directly in R with: install.packages("cxreg") Published software paper The estimation algorithms and software implementation underlying cxreg are now described in the published Journal of Open Source Software article: Younghoon Kim, Navonil Deb, and Sumanta Basu (2026). cxreg: An R Package for Complex-Valued Lasso and Graphical Lasso. Journal of Open Source Software, 11(124), 9698. https://doi.org/10.21105/joss.09698 Key features Complex-valued lasso regression via classo() Complex-valued graphical lasso via cglasso() Full spectral inference pipeline: bandwidth selection, debiasing, variance estimation, confidence regions, and FDR-controlled testing Efficient pathwise coordinate descent for complex-valued data Parallel cross-validation via foreach Functions for transforming complex-valued problems into equivalent real-valued formulations Computational shortcuts including warm starts and active-set screening Example datasets, function documentation, and inference vignette User interface consistent with the glmnet R package Available directly from CRAN Citation Younghoon Kim, Navonil Deb, and Sumanta Basu (2026). cxreg: An R Package for Complex-Valued Lasso and Graphical Lasso. Journal of Open Source Software, 11(124), 9698. https://doi.org/10.21105/joss.09698 Navonil Deb, Younghoon Kim, and Sumanta Basu (2026). Inference for High-Dimensional Sparse Spectral Precision Matrices. arXiv:2606.07986. https://arxiv.org/abs/2606.07986 Navonil Deb, Amy Kuceyeski, and Sumanta Basu (2024). Regularized Estimation of Sparse Spectral Precision Matrices. arXiv:2401.11128. https://arxiv.org/abs/2401.11128 CRAN: https://CRAN.R-project.org/package=cxreg Full changelog: https://github.com/yk748/cxreg/commits/v1.1.4