SimplexRegression 0.1.6
New features
- Added
ic(), a function to compute information criteria
(AIC, SIC, HQIC) and their finite-sample corrections (AICc, SICc, HQICc)
for simplexregression model fits.
Bug fixes,
validation, and breaking changes
- Renamed
penalized.ic() to ic.penalized()
and penalized.ss() to ss.penalized().
- Removed the custom
nobs(), fitted(),
formula(), AIC(), BIC(),
df.residual(), deviance(),
lrtest(), and HQIC() methods for
simplexregression objects; these generics now dispatch
directly to their stats/lmtest defaults. The
Hannan-Quinn criterion is now obtained via ic().
SimplexRegression 0.1.5
New features
- Added the
AISRowing dataset.
- Added
r2(), a function to compute pseudo-R² for
simplexregression model fits.
Bug fixes,
validation, and breaking changes
- Added validation of
lambda > 0 in
parametric_mean_links.
- Added validation of
kappa > 0 in
penalized.ss() and penalized.ic().
- Added error handling in functions that use
solve().
- Removed internal
par() calls from
plot.simplexregression(), halfnormal.plot(),
local.influence(), diag.im() and
diag.distances(); these functions no longer override the
user’s graphical parameters. As a result, the reset.par
argument in plot.simplexregression() was removed, as it is
no longer needed.
Documentation
- Standardized and shortened variable names in the
Biomass, RelativeHumidity, and
AbortionOpposition datasets for consistency across package
datasets.
Dependencies
- Added
moments and parallel to Imports in
DESCRIPTION.
SimplexRegression 0.1.4
- Fixed
opt$counts handling in
simplexreg.fit().
- Added a
digits argument to penalized.ss()
and penalized.ic().
plot.simplexregression() now accepts a cutoff/threshold
argument for flagging observations in residual plots.- Added the
AbortionOpposition dataset.
- Standardized variable names across package datasets.
SimplexRegression 0.1.3