
Research Area: Probability & Statistics
Research Profile: Computational statistics
My research focuses on the applied uses of computationally intensive methods in statistics. One area in which computaional power is particularly useful is in the use of Monte Carlo approximations which can replace intractable calculations or asymptotic argeuments. The use of these methods, however, require reliable software and good diagnostics for when there may be problems with the underling assumptions. In my doctoral work I examined diagnostics for Markov chain Monte Carlo methods which are particularly important in Bayesian statistics. Since then I have worked on the bootstrap and other resampling methods. A suite of functions that I have developed to implement many of these methods is used worldwide. These functions allow for the application of resampling methods in many fields. Much remains to be done, however. I am currently working on applied graphical diagnostics for bootstrap failure. I am also examining new areas of application for these methods and improvements on existing methods. My latest area of research is in Monte Carlo implementations of the EM algorithm for likelihood inference when more standard methods cannot be used.
Computational statistics
2022/2023
Stats 4CI3/6CI3
Stats 771
2021/2022
Stats 2MB3
Stats 4C03/6C03
Stats 771
2020/2021
Stats 4C03/6C03
Stats 752
Stats 771
2019/2020
Stats 2MB3
Stats 4C03/6C03
Stats 743
2018/2019
Stats 2MB3
Stats 3D03
Stats 4CI3/6CI3
Stats 770
2017/2018
Stats 2MB3
Stats 3A03
Stats 743
2016/2017
On Research Leave
2015/2016
Stats 3A03
Stats 4CI3/6CI3
Stats 743
2014/2015
Stats 3J04/3Y03
Stats 4M03/6M03
Stats 743
2013/2014
Stats 3A03
Stats 4CI3/6CI3
Stats 743