Introduction

gamlss is a R package implementing the Generalised Additive Models for Location Scale ans Shape (GAMLSS). The package gamlss is basesd on ideas developed in Rigby and Stasinopoulos (2005), Appl. Statist., 54, pp. 507-554.

There are also three book available;

1. “Flexible Regression and Smoothing: Using GAMLSS in R”; explaining how the GAMLSS models in R can be used in R.

2. “Distributions for modeling location, scale and shape: Using GAMLSS in R”; explaining the explicit and generated distributions available in the package gamlss.dist.

3. “Generized Additive Models for Location Scale and Shape: A distributional regression approach with applications” explaining the different method for fitting GAMLSS i.e. penalised Likelihood, Bayesian and Boosting.

More information about books and papers related to GAMLSS can be found in https://www.gamlss.com/.

The GitHub repository is now hosted under the new gamlss-dev organization: https://github.com/gamlss-dev/gamlss/.

Version 5.4-5

• the dependency of the function chooseDist() on the package pacman was removed (thanks to Henrik Bengtsson for report it)

• the name of the file “fitTail.R” has changed to “loglogSurv.R”

• In the function pbz() we have added an extra check for the option control which check whether the control is a true pbz.control (Tim Cole’s suggestion).

Version 5.4-23

• Tim Cole’s suggestion in predictAll() is added. This is to deal with the problem when mu is fixed.

• Tim Cole’s suggestion in summary()is added. This to fix the problem when y~0, (that is, when there are no df’s), to be incorporated in the summary.gamlss().

Version 5.4-21

• predict() do not print the message “new prediction”

• stepGAIC() produce less lines in the output

Version 5.4-20

• The package is now hosted on GitHub at https://github.com/gamlss-dev/gamlss/.

• Add a new prodist() method for extracting fitted (in-sample) or predicted (out-of-sample) probability distributions from gamlss models (contributed by Achim Zeileis). This enables the workflow from the distributions3 package for all distributions provided by gamlss.dist. The idea is that the distributions3 objects encapsulate all information needed to obtain moments (mean, variance, etc.), probabilities, quantiles, etc. with a unified interface. See the useR! 2022 presentation by Zeileis, Lang, and Hayes for an overview.

Version 5.5-1

• the functions Rsq() and GAIC() have become generic function so they can be used in conjunction with the package gamlss2

• the functions GAIC.scaled() and GAIC.table() have been renamed as GAIC_scaled() and GAIC_table(), respectively, so the name do not class with the generic function GAIC().

Version 5.5-3

• function LR_test is modified to allow ‘gamlss2’ objects.

• the functions GAIC() and Rsq() are now generic function