- Professor of Statistics (Statistics)
Identifying the nature of relationships between variables forms a key part of scientific investigation and regression modelling, where a response variable is related to potential explanatory variables, is an invaluable tool in a very wide variety of settings where measurements of the response variable are subject to a degree of uncertainty. Flexibility can be introduced by freeing the form of relationships from the constraints of linear and other simple patterns and many approaches to this are available. Major themes of my recent research in this area lie in two directions. One is in the development of inferential tools for flexible models, to allow the evidence for the presence and nature of effects to be assessed quantitatively and expressed graphically. The other is in the application of these methods to environmental data, mainly involving air pollution and water quality, to identify spatial and temporal trends which lie at the heart of the very importnat activity of environmental monitoring.
A second major research interest is in statistical shape analysis based on high resolution measurement of surfaces in three-dimensions. There is particular interest in modelling the shape of the human face, in order to quantify any remaining unusual features present after facial surgery or to investigate characteristic associated with particular medical conditions. Methods for formulating the information presented in a facial scan and using this to provide statistical tools for analysis are the focus of the recently formed Face3D consortium.