Skip to content

Latest commit

 

History

History
72 lines (51 loc) · 3.27 KB

NEWS.md

File metadata and controls

72 lines (51 loc) · 3.27 KB
Raw

dominanceanalysis (development version)

dominanceanalysis 2.1.0

  • Support only for R >= 4.0.0
  • Updated README
  • Added support for ordinal regression, using clm

dominanceanalysis 2.0.0

Important API Change:

  • All submodels in fit functions are now calculated using update(), ensuring that all settings of the original model are used in the dominance analysis. If R cannot find the original dataset for update, provide the dataset using the newdata parameter in the dominanceanalysis() method.

Minor Updates:

  • Minor modification in dominanceMatrix.
  • Added complete_flipped_axis parameter for the plot() function. Included examples of flipped axis for plots.
  • Tests for bootstrap analysis (betareg, lmm, glm).
  • Added Nakagawa's indexes for linear mixed models. Refer to README for examples.
  • In the beta regression test, we check that using different estimators generates different results.

Bug Fixes:

  • On R-devel, corrected the number of controls provided for the esoph dataset, fixing a test based on that example.
  • Replaced sort.matrix with .sort.matrix to prevent sorting a standard matrix when devtools are used.
  • Fixed documentation error in plot.dominanceAnalysis: the default plot is for general dominance, not for complete dominance.

dominanceanalysis 1.3.0

  • Added plot() method for dominanceAnalysis objects.
  • Added support for beta regression using the betareg package. Suggested by Maartje Hidalgo.
  • Added 'ordered' parameter to the dominanceMatrix() method.
  • Bug fix: Corrected errors in the logistic regression vignette from v1.2.0 and added some plots for clarification.

dominanceanalysis 1.2.0

  • Added documentation for da.dynlm.fit.
  • Test for the esoph example.
  • New Acknowledgments section in the README file.
  • Removed dependencies on caTools and heplots, as they are only used in examples.
  • Bug fix: Fixed issue where logistic regression indexes failed in some models. Null model likelihood is now correctly calculated using update(x, ~1). Reported by Daniel Schlaepfer.
  • Added basic support for dynamic linear models using the dynlm package. Suggested by Xiong Zhang.

dominanceanalysis 1.1.0

  • Support for named terms in bootDominanceAnalysis and bootAverageDominanceAnalysis.
  • Added dominanceBriefing().

dominanceanalysis 1.0.0

  • First official version. Code coverage of 99% and complete documentation for all methods.
  • Added vignette "Exploring Predictors' Importance in Binomial Logistic Regressions" by Filipa Coutinho Soares.
  • Introduced dominanceMatrix() as a generic for matrix, data.frame, and dominanceAnalysis methods.
  • New retrieval methods for the dominanceAnalysis object: getFits(), contributionAverage(), and contributionByLevel().
  • Removed support for nlme (it never worked well).

dominanceanalysis 0.1.2

  • Added terms parameter in dominanceAnalysis to manually define a set of variables.
  • Updated documentation, thanks to Filipa Coutinho Soares.
  • Test for all relevant functions.

dominanceanalysis 0.1.1

  • Allowed multivariate regression with a covariance matrix.
  • Added documentation for multivariate methods.
  • Added bootstrap analysis for average contribution.
  • Resolved bugs in bootstrap analysis and glm.da.fit.

dominanceanalysis 0.1.0

  • Complete support for OLS, logistic regression, and HLM.
  • Support for bootstrap analysis.