# STAT423 Bayesian Modelling

Methodological and computational aspects of Bayesian statistics, with applications.

Bayesian modelling with an emphasis on scientific data analysis and computing. Topics include Markov chain Monte Carlo, prior choice, posterior assessment, hierarchical modelling, model fitting, and model selection using R, JAGS and other freely available software.

Paper title Bayesian Modelling STAT423 Statistics 0.1667 20 points Semester 2 (On campus) \$1,206.91 Tuition Fees for international students are elsewhere on this website.
Prerequisite
STAT 401 or (STAT 260 and STAT 270 and STAT 310), or equivalent (contact department for further information)
Restriction
STAT 371
Eligibility

Suitable for graduates and professionals interested in advanced modelling with some experience in R programming and knowledge of probability theory.

Contact

Dr Peter Dillingham peter.dillingham@otago.ac.nz

Teaching staff
Paper Structure
• Theory of Bayesian inference.
• Markov chain Monte Carlo methods.
• Modelling using JAGS and Stan.
Teaching Arrangements

This paper is taught via a combination of interactive lectures, hands-on practicals, and project meetings.

interdisciplinary perspective, lifelong learning, and scholarship; and attributes sought by employers communication, critical thinking; ethics; information literacy; research, and self-motivation

Learning Outcomes

Students who successfully complete the paper will:

1. Explain key concepts in Bayesian statistics such as the link between the likelihood, prior and posterior distributions.
2. Understand the relationship between Bayesian and frequentist approaches.
3. Understand sufficient theory to find analytical solutions to standard problems (note: ‘standard’ problems in Bayesian statistics require advanced statistical knowledge).
4. Be able to independently use R with JAGS or stan to complete Bayesian statistical analyses, including the ability to correctly format and manipulate input different data types, run analyses and diagnostics, and interpret and plot results.
5. Be able to implement their own Markov chain Monte Carlo samplers in R.
6. Communicate results to others and understand the ethical and scientific importance of reproducible research.
7. Independently develop advanced hierarchical statistical models linked to a scientific study, create and perform an appropriate Bayesian analysis.

## Timetable

### Semester 2

Location
Dunedin
Teaching method
This paper is taught On Campus
Learning management system
Other

#### Lecture

Stream Days Times Weeks
Attend
A1 Tuesday 09:00-09:50 28-34, 36-41
Wednesday 09:00-09:50 28-34, 36-41
Thursday 09:00-09:50 28-34, 36-41

#### Tutorial

Stream Days Times Weeks
Attend
A1 Thursday 15:00-16:50 28-34, 36-41