The ACT branch proudly offers an in-person workshop: Introduction to Distributional Regression presented by Professor Gillian Heller.
Occasionally courses have to be cancelled due to a lack of subscription. Also, places are limited, so make sure you register early to guarantee you spot and ensures that the above will not happen! Early Bird Rego closes 10 August 2025.
Generalized Additive Models for Location, Scale and Shape (GAMLSS), introduced 20 years ago by Rigby and Stasinopoulos, are a very general framework for univariate regression. Their novelty arises from the fact that all the parameters of the assumed parametric response distribution can be modelled as functions of the explanatory variables, enabling modelling of the complete distribution rather than just the mean. The assumed response distribution can in principle be any computable distribution. Methods for modelling the distribution parameters using explanatory variables include linear terms, smoothing terms, and any sensible machine learning technique such as neural networks, LASSO, principal component regression etc. There are also different ways of fitting the model, which include the classical (using penalized likelihood), Bayesian, or boosting methods.
This workshop is designed for practitioners and applied statisticians who would like to learn how to model their datasets using GAMLSS. We will discuss the distributional regression framework, selection of the response distribution, and selection of explanatory variables modelling each parameter of the response distribution. Model diagnostics, interpretation, and the use of the GAMLSS model in prediction will also be discussed.
Bio:
Professor Gillian Heller is a biostatistician with an interest in distributional regression. Gillian is Professor of Biostatistics at the NHMRC Clinical Trials Centre, University of Sydney. Gilian has coauthored four books in the regression modelling area, the first directed towards actuarial applications of the generalized linear model, and the remaining three focusing on distributional regression, in particular the GAMLSS framework.
Requirements:
This course is designed for individuals who wish to learn how to perform distributional regression modelling using the gamlss software. We assume users have the following background/experience:
* Familiarity with R
* Experience using linear models and also, preferably, generalized linear models
We recommend installing the following R packages on your laptop/machine ahead of time:
* gamlss (from CRAN)
* gamlss2 from GitHub, installed using install.packages("gamlss2", repos = c("https://gamlss-dev.R-universe.dev", "https://cloud.R-project.org"))
Course materials will be made available online before the start of the course in a timely manner.
After this course, you will be able to:
1. Fit distributional regression models in gamlss and gamlss2;
2. Check model fit using appropriate diagnostic tools;
3. Interpret results of gamlss models, using visualisations of the model outputs for use in understanding model fit and for publication.
Registration for the workshop includes morning tea, lunch, and afternoon tea.
Parking at the workshop venue can be arranged. If requires, please contact ssacanberra@gmail.com
Cancellation policy:
Cancellations received prior to 11:59pm Canberra time 3 September 2025 will be refunded, minus the Stripe processing fee (1.75% + $0.30 per transaction) and an SSA administration fee of $20. From then onwards, no part of the registration fee will be refunded. However, registrations are transferable within the same organisation. Please advise any changes to events@statsoc.org.au.
If you have any questions, please contact ssacanberra@gmail.com or events@statsoc.org.au
@StatSocAus
