RIFanalysis 0.9.3
- Made the power-law reference test portable to R builds without
long-double support by allowing the observed one-unit difference in the
final decimal of the bootstrap alpha standard deviation. This change
affects only the test tolerance; the statistical implementation is
unchanged.
- Reduced the number of bootstrap simulations in the executable
rif_compute() documentation example from 50 to 10, keeping
the example representative while providing sufficient margin below
CRAN’s requested five-second execution limit.
RIFanalysis 0.9.2
- Added a namespace-safe, reproducible manual bootstrap engine and
made it the default across the fitting and workflow APIs.
- Deprecated
bootstrap_engine = "poweRlaw" and
engine = "poweRlaw"; these options are retained for
compatibility and are planned for removal in version 1.0.0.
- Standardized the Z1, Z2, and Z3 workflow arguments on
save_plots and plot_dpi.
- Deprecated the Z3 arguments
save and dpi
in favor of save_plots and plot_dpi; the
aliases are planned for removal in version 1.0.0.
- Updated the Z3 workflow so that plots are not saved by default
(
save_plots = FALSE) and renamed its plot export argument
from formats to plot_formats, consistently
with the Z1 and Z2 workflows.
- Increased the default PNG export resolution to 600 DPI.
- Added complete Excel exports for power-law analyses. When Z2 is run
from prepared data, it now produces an analysis workbook for each input
group containing the original data, MLE-KS estimates, bootstrap results
and p-values, the combined MLE-KS/bootstrap summary, and the individual
bootstrap simulations.
- Extended and tested these Excel exports for datasets both with a
single default type and with multiple values in the
Type
variable; bootstrap simulations from multi-type analyses are identified
by type in the exported workbook.
- Made the package vignettes self-contained by using the example
datasets distributed with the package.
- Normalized the multi-type example data to portable ASCII text.