NEWS.md

In bpnreg: Bayesian Projected Normal Regression Models for Circular Data

bpnreg 2.0.3

• Small documentation fixes for the breaking change:

Using @docType package no longer automatically adds a -package alias. Instead document _PACKAGE to get all the defaults for package documentation.

• Documentation fix for unused arguments in b_samp and betaBlock

• Update JSS hyperlinks

bpnreg 2.0.2

Bug fixes

• Made sure that formulae for SAM, AS and bc correspond to those in Cremers, Mulder \& Klugkist (2018). Circular interpretation of regression coefficients. In previous versions of the package there was an error in the computation of SAM, AS and bc causing erroneous results.

• Corrected c++ code of mcmc sampler for intercept-only circular mixed effects model to solve error warning in v.2.0.0 and v.2.0.1.

• Corrected incorrect computation of model fit estimates for circular mixed effects models in v.2.0.0 and v.2.0.1.

• Made sure that the chosen lag is displayed correctly in the output.

bpnreg 2.0.1

Bug fixes

• Removed MASS from imports to fix note in the 2.0.0 version

• Changes in c++ code to solve errors from CRAN valgrind checks in the 2.0.0 version.

• Changed arma::inv_sympd to arma::inv in c++ code to solve errors on macOS in the 2.0.0 version.

bpnreg 2.0.0

Major changes

• Changed the names of the mcmc output objects. B1, B2, Beta.I, Beta.II, B.I, B.II, VCov.I and VCov.II are now named beta1, beta2, beta1, beta2, b1, b2, omega1 and omega2 respectively.

• Functions predict() and residual() were removed.

• MCMC sampler underlying the bpnme function is rewritten in c++. This should significantly increase computational speed.

• Increased memory efficiency by not storing intermediate objects in memory.

• A vignette containing answers to frequently asked questions about usage of the package is included.

Bug fixes

• Adapted documentation bpnme function such that it clearly states only one grouping factor can be used.

• Made sure that the circular random intercept variance (and circular random slope variance of categorical predictors) is computed as 1 - mean resultant length. Also updated documentation accordingly.

• Results for bc, AS and SAM are now also included in degrees if units = "degrees" in circ_coef function

• An error warning in case of any missing values was included.

• The WAIC criterion in the model fit estimates obtained from the fit() function is more correctly referred to as WAIC1.

• Format of cRS object stored in a bpnme fit object was changed from matrix to data.frame to accommodate for data of more than two types, e.g. character and numeric.

bpnreg 1.0.3

• Deleted arguments from function documentation that are not in \usage.

• Updated references in the packages description.

bpnreg 1.0.2

• This version contains an update of code containing the class function to make sure the package works with R version 4.0.0

• The check for the range of the circular outcome was corrected.

bpnreg 1.0.1

• The function bpnme can now also handle the standard R data.frame instead of only Data Frame Tbl (from dplyr, tbl_df(data))

• The function coef.bpnr now correctly returns radians instead of degrees when option units == "radians" is chosen.

• The documentation now states that the circular outcome needs to be measured in radians.

• The references in the documentation have been updated.

• The functions mmr and mmme now give out a warning message when the circular outcome is out of range.

• An error in the assignment of column names in the random effect model matrix (in mmme function) was fixed.

• An error in the check for whether random effects are numeric (in bpnme function) was fixed.

• Several checks and error warnings for wrong specifications of nesting structure were added.

• Internal functions were taken out of the package documentation on CRAN.

• The priors for the (fixed effect) regression coefficients in the regression and mixed-effects model are now both N(0, 10000), We also note this in the documentation of bpnr() and bpnme().

• We note the fact that the mixed-effects model is only developed for models with one nesting variable in the documentation.

bpnreg 1.0.0

• This is the first version of the package.

bpnreg documentation built on May 29, 2024, 12:35 p.m.