Paddy is in cohort 10 and graduated from University College Dublin in 2020 with a BSc in Statistics. Since then, he has worked as a Data Science Intern at a large Petrochemicals company, as a Statistician with a Health Data Analytics startup, and as a Research Assistant within the HIV Inference Group at Imperial College London, where he worked on modelling male circumcision coverage in sub-Saharan Africa. Paddy is mainly interested in Bayesian and spatio-temporal statistics and machine learning, particularly within epidemiological and environmental applications. He is also interested in programming and scientific computing. Outside of academia, he plays hurling and Gaelic football, and also enjoys reading, films and a nice pint.

Project title:
Developing clustering algorithms for conditional extremes models

Supervisor(s):
Christian Rohrbeck, Jordan Richards (University of Edinburgh), Vangelis Evangelou

Project description:
Conditional extreme value models have proven useful for analysing the joint tail behaviour of random vectors. While an extensive amount of work to estimate conditional extremes models exists in multivariate and spatial applications, the prospect of clustering for models of this type has not yet been explored.  This PhD project aims to develop clustering algorithms to analyse conditional extremes models in a wide range of settings. These clustering methods will be designed to aid exploratory analysis and/or parameter estimation. The methods will be implemented to test their computational cost and performance will be assessed rigorously using simulated and real-world data. Possible applications include the analysis of several weather variables across multiple spatial sites.

References: Janet E. Heffernan and Jonathan A. Tawn. A conditional approach for multivariate extreme values (with discussion). Journal of the Royal Statistical Society Series B: Statistical Methodology, 66(3):497546, July 2004. ISSN 1467-9868. doi: 10.1111/j.1467-9868.2004.02050.x.