Overview
Topic status: We're looking for students to study this topic.
Approximate Bayesian Computation (ABC) is a new development which can be used for Bayesian inference when the likelihood is intractable but simulation of data from the likelihood is relatively easy. ABC replaces the likelihood by repeated sampling of data and demanding closeness to the observed data in terms of summary statistics. Applications have been various such as in finance, statistical genetics, biological networks and disease transmission modelling. A variety of Monte Carlo algorithms, such as Markov chain Monte Carlo and Sequential Monte Carlo, can be used for inference. These approaches are very computationally intensive. Efficiency and being self-adaptive are important properties and we have developed fully adaptive algorithms. However these algorithms can be improved further. The project will review the literature, carry out experiments to investigate the properties of suggested improvements to the currently available algorithms and apply the approach to an example involving infectious diseases transmission.
- Study level
- PhD
- Supervisors
- QUT
- Organisational unit
Science and Engineering Faculty
- Research area
- Contact
-
Please contact the supervisor.
Professor Tony Pettitt
