Introduction to Bayesian statistics BIOL5124
- Academic Session: 2019-20
- School: Biodiversity Animal Health Comp Med
- Credits: 10
- Level: Level 5 (SCQF level 11)
- Typically Offered: Semester 2
- Available to Visiting Students: Yes
- Available to Erasmus Students: No
This course will provide a basic introduction to Bayesian statistical philosophy and theory, and a more detailed introduction to the fitting of statistical models using Markov Chain Monte Carlo techniques in WinBugs and JAGs.
Concentrated course offered over one week, with 1 hour lecture and 2 hours computer laboratory per day.
Requirements of Entry
A second class Honours degree or equivalent in a relevant subject. Professional experience may be taken into account.
If your first language is not English, the University sets a minimum English Language proficiency level. See English Language Requirements
Student must have undertaken course BIOL5133 Programming in R
Students will submit practical exercises to gauge their depth of understanding and engagement with the skills learned in each of the practical sessions. The work will be assessed not only on completion of the assigned tasks but on interpretation and self-reflection of the theories learned (50%). The remaining 50% will be a take-home problem-based assignment that will require integration of the knowledge and skills learned in this module, in the analysis and discussion of an independent dataset.
The aim of the course is to provide the student with an evidence-based founding in the basic theory and practice of Bayesian statistics.
Intended Learning Outcomes of Course
By the end of this course students will be able to critically discuss with reference to theory and practice:
■ The key differences between a Bayesian and frequentist approach
■ How prior information is used in a Bayesian approach
■ The concept of Markov Chain Monte Carlo techniques
■ The distinction between Metropolis-Hastings and Gibbs sampling
In addition, they will be able to:
■ Write simple programs in WinBugs or JAGs
■ Specify and discuss critically the appropriate use of both informative and 'uninformative' priors
■ Identify when a model has converged
■ Conduct model selection using DIC
Minimum Requirement for Award of Credits