NEWS
lmeresampler 0.2.4 (2023-02-11)
- Adding tests for rboot noise feature.
lmeresampler 0.2.3 (2023-01-30)
- There are no user-visible changes. This fixes an error with testing on CRAN based on RNG differences.
- 'rbootnoise' argument added to residual bootstrap to bypass "system is exactly singular" issues with lme4 models (Credit to Ilmari Tamminen)
- Eliminated unnecessary warnings for unnamed vectors with type = "reb2".
- 'bootstrap_pvals()' added mean centering for correction to the calculation
lmeresampler 0.2.2 (2022-04-29)
NEW FEATURES
- Additional auxilliary distributions were added for the Wild bootstrap. There are now 6 options, including the standard normal.
- '.refit' argument can be set to FALSE in order to return the only bootstrap responses.
- 'bootstrap_pvals()' appends bootstrap p-values to the summary table for the fixed effects
- 'combine_pvals()' provides a way to combine the results of 'bootstrap_pvals()' for parallel runs.
BUG FIXES
- 'plot.lmeresamp()' now works if the replicates are a numeric vector rather than a data frame or tibble.
- bug fixed when 'na.action = na.omit'
- fixed issue with transformed variables in 'glmer'
lmeresampler 0.2.1 (2022-03-23)
- Unarchiving from CRAN
DEPENDENCY CHANGE
- Remove 'catchr' dependency to avoid issues on CRAN
BUG FIXES
- message/error/warning summarization in 'summary.lmeresamp' has been fixed
- If 'var' is omitted from 'plot.lmeresamp()' a halfeye plot with all terms is created.
lmeresampler 0.2.0 (2021-05-01)
- The case, parametric, and residual bootstraps now suppport 'glmerMod' objects.
- The Wild bootstrap is available for 'lme' and 'lmerMod' objects.
- The CGR bootstrap is now the default "residual" bootstrap algorithm.
- Objects returned by the 'bootstrap()' call are now of class 'lmeresamp'.
- 'lmeresamp' objects have a new structure, including a new 'stats' dataframe (contains the observed value, bootstrap mean, standard error, and bias of each LME model parameter).
- New generic 'print()' function that is compatible with 'lmeresamp' objects
- New generic 'confint()' function that is compatible with 'lmeresamp' objects (the possible confidence intervals include: basic, normal, percentile, or all)
- A package vignette is now available
- Vignette outlines how to perform parallelization in 'bootstrap()' using the 'doParallel' and 'foreach' packages
- New 'combine()' function that combines processes split for parallelization for unified output
lmeresampler 0.1.1 (2020-01-31)
- Unarchiving from CRAN - back to active development
- Updating for use with the new version of dplyr (>= 0.8.0)
- Bug fixed for 'case_bootstrap.lme ' so that '.cases.resamp' can be found
lmeresampler 0.1.0 (2016-07-15)