Practical Bayes for Beginners
Presenters: Prof. Kerrie Mengersen and Dr. Erin Peterson
Overview
Bayesian modelling and data analysis are becoming a standard part of the statistical toolkit. Its appeal includes the availability of hierarchical models for better describing complex systems, the use of priors to describe uncertainty and include external information in the analysis, and the direct probabilistic interpretation of the results.
While simple Bayesian models can be analysed analytically, most analysis is via Monte Carlo methods such as Markov chain Monte Carlo (MCMC). There is a great range of MCMC algorithms available now for Bayesian computation.
This one-day course introduces the practicing statistician to Bayesian analysis. The course is strongly practical, with emphasis on understanding the fundamental concepts, modelling in a Bayesian context, using MCMC and ‘doing’ Bayesian analysis via the software package R.
Please note that this course is introductory. It assumes some knowledge of statistics but no knowledge of Bayesian or MCMC approaches.
IT Requirements
Participants are requested to bring a laptop with the following software loaded:
- R (freeware statistics program)
Course Outline
The following topics will be covered:
- What is Bayesian Statistics
- Priors, models & results
- Common MCMC algorithms
- Model fit & model choice
- Role & formulation of priors
- Examples of different types of models
- Reporting of Bayesian analysis results, with examples from published literature
Target Audience
Practicing statisticians wanting to learn and apply the fundamentals of Bayesian analysis or researchers in other disciplines with a statistical knowledge equivalent to 2 years of undergraduate study.
Basic statistical knowledge, and statistical computing, but no knowledge of Bayesian methods.
Learning Objectives
Attendees will gain a basic understanding the fundamental concepts, modelling in a Bayesian context, using MCMC and ‘doing’ Bayesian analysis via the software package R. Participants will be introduced to a range of models for describing complex data and the application of these models to real problems in environmental and agricultural science, as well as ecology.
Presenter Profiles
Kerrie Mengersen has held a Chair in Statistics at Queensland University of Technology (QUT) for over 15 years. She is one of Australia’s most prominent Bayesian statisticians, is currently an Australian Research Council (ARC) Laureate Fellow (2015-2020), and is the Deputy Director of the ARC Centre of Excellence for Mathematical and Statistical Frontiers (ACEMS) in Big Data, Big Models, and New Insights. Kerrie is an Accredited Member of the Statistical Society of Australia (2001) and Society President, a long-time executive member of the International Society for Bayesian Analysis, an elected Fellow of the Royal Statistical Society (2004) and an elected Fellow of the Institute for Mathematical Sciences (2005). She has over 200 refereed journal publications and sustains a Bayesian Research and Applications Group (BRAG) comprising postgraduate students and early career researchers working on a variety of methodological and applied problems in health, environment and industry.
Erin Peterson is a Principle Research Fellow in the Institute for Future Environments and ACEMS at QUT. Her educational background and experience allow her to work in environmental statistics, geographic information science, ecology, and environmental monitoring and her work has been published top-tier journals, including the Journal of the American Statistical Association and Ecology Letters. She was awarded the CSIRO Julius Career award (2014-2016) based on the scientific impacts of her research, but was also part of a multidisciplinary team awarded the 2016 US Forest Service Rise to the Future Award for the successful knowledge transfer of statistical methods that have become standard-practice for aquatic resource management. Erin is currently a Regional Representative of The International Environmetrics Society (TIES).
Date: 17 July, 2016
Maximum Number of Participants: 40