Matthew Pawley

About me

I am a statistics PhD student at the University of Bath as part of the SAMBa CDT.

My research develops statistical learning methodology for multivariate extreme value analysis. My supervisors are Dr Christian Rohrbeck and Dr Evangelos Evangelou.

I completed an MSci Mathematics degree at Imperial College London, graduating with First Class Honours in 2020. My master's project developed a deep learning model for survival analysis and was supervised by Professor Axel Gandy. I achieved a Distinction in MRes Statistical Applied Mathematics at the University of Bath in 2021.

My PhD project

Title: Extensions and Applications of the TailPairwise Dependence Matrix
Supervisors: Dr Christian Rohrbeck, Dr Evangelos Evangelou

My research concerns modelling extremal dependence in high dimensions using the tail pairwise dependence matrix (TPDM). My thesis addresses the following topics:

1. Testing for changes in extremal dependence.
2. Connections to compositional data analysis (CoDA).
3. Inference for max-linear models and estimation of failure probabilities.
4. Bias-corrected estimation of the TPDM.

Keywords: extreme value theory, extremal dependence, tail pairwise dependence matrix, statistical learning, dimension reduction

Publications

1. Elsom, H. and Pawley, M. (2024)
   Extreme Value Statistics for Analysing Simulated Environmental Extremes
   To appear in Extremes -- arXiv preprint available
   

Talks

1. Testing for time-varying extremal dependence
   SAMBa Summer Conference 2024
   10 July 2024
   University of Bath, UK
2. Testing for changing extremal dependence in high dimensions
   Lancaster Extremes Reading Group
   04 June 2024
   Lancaster University, UK
3. Multivariate extreme value statistics for analysing simulated environmental extremes
   Bath Statistics Seminar
   05 March 2024
   University of Bath, UK
4. Estimation of the tail pairwise dependence matrix
   STOR-i Extremes Workshop (STEW) 2023
   20 September 2023
   Lancaster University, UK
5. Estimating the probability of very rare events
   SAMBa Summer Conference 2023
   11 July 2023
   University of Bath, UK
6. Sparse robust principal component analysis for extremes
   Extreme Value Analysis (EVA) 2023
   29 June 2023
   Bocconi University, Italy
7. Sparse robust principal component analysis for extremes
   Turin-Bath PhD Workshop in Applied Probability and Statistics
   20 June 2023
   University of Torino, Italy
8. Shrinkage estimation of the tail pairwise dependence matrix
   KAUST Extremes Seminar
   21 November 2022
   KAUST, Saudi Arabia (virtually)
9. Statistical learning for multivariate extremes
   Research Students' Conference in Probability and Statistics
   07 September 2022
   University of Nottingham, UK
10. Extremal dependence: what it is and why it matters
    SIAM-IMA Three Minute Thesis Competition (First Prize Winner)
    26 May 2022
    University of Bath, UK

Academic workshops with industrial partners

1. Non-stationarity in extreme sea levels in the UK
   Environment Agency, JBA Trust
   University of Bath, January 2023
2. Discovery of biodiversity, physical and metagenomic patterns from TARA Oceans expeditions
   Centro de Modeliamento Matemático, TARA-CEODOS
   Universidad de Chile, December 2022
3. Broad-band to narrow-band diffraction patterns
   Central Laser Facility
   University of Bath, January 2022
4. Physiological based clustering for arthritis joint progression in space and time
   NHS
   University of Bath, June 2021
5. Joint modelling of progression-free and overall survival for adaptive clinical trials
   Roche
   University of Bath, January 2021
6. Dose-response modelling exploiting longitudinal data
   Novartis
   University of Bath, September 2020

Tutoring

• MA10207 Analysis 1A
• MA10274 Sequences and Functions
• MA10212 Probability & Statistics 1B
• MA20227 Introduction to Data Science
• MA30085 Time Series

Contact

Email: mtp34@[Uni of Bath domain]
Office: 4 West 2.15

Last updated: 12 October 2024