Functional data analysis with informative missingness (UK Only)

This PhD research project focuses on developing functional principal component analysis (FPCA) and multivariate FPCA (MFPCA) frameworks for functional data with informative missingness. The increasing availability of extensive datasets with variables measured over a continuum has opened new frontiers in research across disciplines. These datasets, known as functional data, are prevalent in fields ranging from environmental monitoring and education to biomedical sciences and engineering. However, analysing such data remains challenging due to their high dimensionality, multivariate structure, and critically, the presence of missingness that is not completely at random. This project aims to fill the gap by developing advanced statistical methods to handle informative missingness in functional data analysis. Real-world datasets often suffer from different types of informative missingness, and applying FPCA or MFPCA without accounting for these issues can lead to biased results, misrepresenting relationships among variables and undermining downstream analyses such as prediction, classification, and decision-making. The application will be within biomedical studies, such as Alzheimer's disease and scleroderma. The position offers a highly competitive EPSRC Doctoral Landscape Award providing full academic fees, together with a tax-free maintenance grant at the standard UKRI rate (£20,780 in academic session 2025/26) for 3.5 years. Training and support will also be provided, and candidates will be automatically considered for a School of Mathematics Scholarship.