Bachelor of Mathematics (Statistics), QUT
- Bayesian computational statistics
- Markov chain Monte Carlo (MCMC)
- Likelihood-free methods
- Pseudo-marginal methods
The introduction of Markov chain Monte Carlo (MCMC) sampling methods to the statistical community in the 1990s meant major advances in the capacity to perform statistical analyses in a Bayesian framework for non-trivial applications. However, when the likelihood function is intractable or other computational issues arise, sampling methods that accommodate more complex models are required. Despite advances in MCMC methodology and the introduction of likelihood-free methods, significant work is required to find improved adaptive and efficient methods for performing inference. My work focuses on alternative sampling methods for complex models, especially pseudo-marginal and likelihood-free methods which are useful when the likelihood function is intractable.
Sequential Monte Carlo for static Bayesian models with independent MCMC proposals
Happy to say that my PhD thesis on “contributions to computational Bayesian statistics” is now available online! https://eprints.qut.edu.au/132155/