UoA 22: Statistics

Research activity in the Department of Statistical Science at UCL covers a wide spectrum, ranging from development of generic statistical theory and methodology to applications in industry, science and medicine.  There has been a significant level of collaborative and interdisciplinary research and the breadth of our research activities is reflected in the variety of our sponsors.  The UCL element of new awards (individual and collaborative) associated with research active staff over the period was over £2.4M, including Leverhulme Trust/ESRC (£970K), DEFRA (x2) (£424K), EPSRC (x6) (£422K), DH (£157K), BBSRC (£136K), Cancer Research (£120K), Garfield Weston Trust (£102K), Leverhulme Trust (£98K), Bowling-Pfizer (£17K), NHS R&D (£10K).

Our research students have been funded by EPSRC, MRC, DEFRA, NMLC, EU, ORS and various foreign governments.  Over 340 refereed journal papers (published from UCL) appeared in statistical and mathematical journals, including Ann. Statist., Biometrika, J.& A. Appl. Prob, J. R. Statist. Soc., Proc. Roy. Soc., Scand. J. Statist., Statistics in Medicine, and in numerous diverse outlets such as BMJ, Circulation, Forensic Science International, IEEE Trans. Signal Processing, J. Energy Engineering, J. Hydrology, The Lancet, Theoretical Population Biology.  The statistical material in the latter category of papers ranged from original methodology and theory to substantive contributions to the discipline.  Members of the Department (while at UCL) gave over 50 invited lectures at major international conferences and workshops and numerous presentations and seminars nationally.

Our research mission is to “identify, develop and apply the sound statistical foundations essential for advances in the physical, social, environmental and medical sciences”.  We aim to be world leading in foundations of statistics and generic theory and methodology, while making important contributions to key application areas.

Research in the Department is grouped under four broad themes (theme
leader starred, Early Career Researchers in italics):

  • Foundations and Theory (Sweeting*, Dawid, Didelez) covers foundational aspects of probability and inference, particularly Bayesian theory, decision theory, predictive inference, probabilistic expert systems, graphical modelling, causal inference and event history analysis.  The Leverhulme-funded interdisciplinary research programme Evidence, Inference and Enquiry (worth just under £1M) incorporates researchers from Economics, History, Medicine, Psychology, Philosophy, Education, Law and Forensic Science.

  • Applied Probability and Stochastic Modelling (Isham*, Chandler, Galbraith, Guillas, Northrop, Wilkinson-Herbots) includes modelling and inference for spatial-temporal processes as a major component, with hydrology and climatology as important applied areas, with DEFRA, EC and the NERC/TSUNAMI consortium providing significant funding (almost £1M).  Further important research concerns the modelling of geological processes, while theoretical research on epidemic models and genetics leads to applications in the life sciences and insight on biological mechanisms.

  • Methodology and Applications (Fearn*, De Reyck, Hennig, Olhede, Siddiqui) includes estimation and inference for stochastic-mechanistic and large spatio-temporal models by the Stochastic Modelling group, adjustment for measurement error in epidemiological studies, signal processing, wavelet representations, image analysis and operational research.  Core methodology for multivariate data features strongly, particularly cluster analysis, multivariate calibration, classification and analysis of high dimensional data.  Other major application areas include mathematical biology, medical imaging, geophysics, chemistry, project portfolio management, energy economics and law.

  • Statistics in Medical Research (Omar*, Ambler, Royston, Seaman, Senn) encompasses both applied health research and the development and evaluation of statistical methodology, with significant funding (c£700K) from DH, NHS R&D, MRC and various charities and pharmaceutical organisations.  Current methodological topics include risk modelling of health outcomes, modelling clustered data, meta-analysis, evidence synthesis and Bayesian methods.  Applied research has included developing and validating risk models, measuring variations in health outcomes, randomised trials, and large scale observational studies.

Download full text of the RA5a statement for Statistics & Operational Research (pdf 96Kb)

Staff names below link to submitted publications:

[external source element is broken]
error message: The database connection Oracle_database_connection2 cannot be found.

Page last modified on 16 jan 08 01:34