Skip to contents
GLBFP 0.5.1
Features
- Added sparse-prefix computation for ASH, LBFP and GLBFP grid-count estimators.
- Added
compute_Di() and compute_di() for fixed-grid leave-one-out self-support scores.
- Added lowercase aliases (
ash(), lbfp(), glbfp() and corresponding grid functions).
- Added
as.data.frame() helpers for grid and D_i score objects.
- Added
plot.glbfp_di() for leave-one-out score diagnostics.
- Added a hexagonal package logo and expanded pkgdown article structure.
- Added test infrastructure (
testthat edition 3) and regression/invariant tests.
- Added S3 helpers for
summary() and predict() on fit/grid objects.
- Added two publication-oriented vignettes (introduction and validation).
- Added reproducible benchmark scripts in
benchmarks/ and validation helpers in inst/bench.
- Added pkgdown and coverage workflows.
- Added citation files (
CITATION.cff, inst/CITATION) and cran-comments.md.
- Added pre-R Journal preparation files in
dev/ and rjournal-prep/.
Fixes
- Strengthened input validation (finite numeric checks, shape checks, explicit errors).
- Improved numerical robustness in
compute_bi_optim() for near-singular covariance matrices.
- Fixed plotting methods for irregular custom grids and metadata handling.
- Fixed broken references in
plot.ASH_estimate().
- Reduced duplicated S3 plotting and printing code through shared internal helpers.
GLBFP 0.4.0
- Added standard deviation and confidence interval output for
*_estimate functions.
GLBFP 0.3.0
- Internal maintenance release.