NEWS

sssvcqr 0.0.4 (2026-05-15)

Addressed CRAN feedback from the 0.0.3 submission. Removed the https://stork343.github.io/sssvcqr/ URL from DESCRIPTION and from README.md because the corresponding pkgdown site has not yet been deployed to GitHub Pages; the URL will be reinstated in a future release once the site is live. Replaced the relative [LICENSE](LICENSE) link in README.md with a plain-text reference to the file shipped with the package.

Updated graph construction to use sparse k-nearest-neighbor matrices.
Enforced degree-weighted centering constraints in the ADMM delta updates through sparse KKT solves.
Added known-truth simulation outputs, selection-recovery summaries, expanded numerical tests, and improved plot colorbars.
Added top-level JSS replication materials for synthetic, blocked-CV, comparison, Lucas County package-sample examples, and full-data Lucas County context checks.
Added a jss-submission/ manuscript snapshot for source-archive review while excluding it from the R package tarball.

Strengthened input validation, fold-wise adaptive-weight estimation in blocked cross-validation, and inverse-distance extrapolation for predict(..., k > 1).
Added a reproducible comparison section for the JSS replication material.

Initial package scaffold for sparse-smooth spatially varying coefficient quantile regression.
Added ADMM fitting, prediction, spatially blocked cross-validation, graph construction, simulation helpers, and KKT diagnostics.
Added tests, vignettes, a Lucas County sample data set, CI configuration, contribution guidelines, and a JOSS paper draft.
Added a standard plot() method for fitted model objects in line with JSS expectations for R packages returning compound objects.