Applied statistics has a rich history at Reading: numerous methodological developments by researchers in this group have given rise to important impacts in a wide range of applications.
Our current research continues in this tradition, with the development of new statistical methods for analysis of large and complex data sets being increasingly important in the current era of 'big data'.
Current research interests are predominantly aligned to four overlapping areas.
In recent years there has been considerable interest in accelerating and refining the drug development process. Statistical innovations in trial design have led to the availability of new approaches to address these objectives. Our group at Reading is primarily involved in the development of methodology within the fields of adaptive designs, in particular adaptive seamless designs & group-sequential trials, pharmacogenetics and pharmacokinetics.
We have strong research links with other academic groups in this field including those based at the Universities of Warwick, Sheffield and Leeds. We also collaborate with both the pharmaceutical industry and public sector institutions. Within our group, several PhD students are studying for the qualification part-time whilst working in the Industry.
People: Sue Todd, Fazil Baksh
Recent developments in whole genome sequencing have led to an explosion in the availability of genetic data. The main challenge is now in drawing inferences from these data. Our group develops methods for achieving this, and uses such methods in applied studies. Areas of particular interests are genome-wide association studies (in both humans and bacteria), family and population genetics for whole genome data, and inference for coalescent models and pathogenomics.
We collaborate with the Modernising Medical Microbiology consortium and East Malling Research, the Infectious Disease Research Centre at Massey University New Zealand, along with the Microbiology, Evolutionary Biology and Crop Research groups at Reading.
People: Fazil Baksh, Richard Everitt
The Bayesian approach to statistical inference has seen major successes in the past twenty years, finding application throughout science, engineering, finance and other disciplines. The main driver of these successes was the development of Monte Carlo computational methods to perform statistical inference. Our group at Reading is playing an active part in the development of next-generation Monte Carlo techniques, to exploit the increasing size and complexity of modern data sets.
Particular areas of expertise are approximate Bayesian computation, sequential Monte Carlo, and approximate and adaptive MCMC methods. Recent applications include genetics, neuroscience, signal processing and network analysis.
We have a regular Reading Group in Bayesian Computation, also involving members of the Data Assimilation, Polymer Physics, Ecology and Evolutionary Biology groups. This is a part of a university-wide network, the 'Bayes Group', of researchers interested in the use of Bayesian statistics.
People: Richard Everitt
Statistics of Extremes
Extreme value theory provides a rigorous and prolific framework for analysing rare events with severe impact. The development of extremes over space and/or time rests on the celebrated max-stable processes. A recent offspring of extreme value theory is the generalised Pareto process.
Our group has recognised contributions in developing statistical tools for rare events with application to environmental data. We have also established strong international ties with other academics in the field.
People: Claudia Neves
+44 (0) 118 378
|Professor Sue Todd||Professor of Medical Statistics||8917||s.c.todd|
|Dr Fazil Baksh||Lecturer||8034||m.f.baksh|
|Dr Richard Everitt||Associate Professor||8030||r.g.everitt|
|Dr Claudia Neves||Lecturer||7931||c.neves|
Professor Robert N Curnow
Our lively research environment is supported by a statistics seminar series and research workshops (with strong links to the Probability and Stochastic Analysis theme), together with a strong presence in the Royal Statistical Society local group.
Everitt, R. G. (2018) Efficient importance sampling in low dimensions using affine arithmetic. Computational Statistics, 33 (1). pp. 1-29. ISSN 1613-9658 doi: https://doi.org/10.1007/s00180-017-0729-z
Dimier, N. and Todd, S. (2017) An investigation into the two-stage meta-analytic copula modelling approach for evaluating time-to-event surrogate endpoints which comprise of one or more events of interest. Pharmaceutical Statistics, 16 (5). pp. 322-333. ISSN 1539-1612 doi: https://doi.org/10.1002/pst.1812
Everitt, R. G., Johansen, A. M., Rowing, E. and Evdemon-Hogan, M. (2017) Bayesian model comparison with un-normalised likelihoods. Statistics and Computing, 27 (2). pp. 403-422. ISSN 1573-1375 doi: https://doi.org/10.1007/s11222-016-9629-2
Kunz, C. U., Stallard, N., Parsons, N., Todd, S. and Friede, T. (2017) Blinded versus unblinded estimation of a correlation coefficient to inform interim design adaptations. Biometrical Journal, 59 (2). pp. 344-357. ISSN 0323-3847 doi: https://doi.org/10.1002/bimj.201500233
Prangle, D., Everitt, R. G. and Kypraios, T. (2017) A rare event approach to high-dimensional Approximate Bayesian computation. Statistics and Computing. ISSN 1573-1375 doi: https://doi.org/10.1007/s11222-017-9764-4
Wollstadt, P., Sellers, K. K., Rudelt, L., Priesemann, V., Hutt, A., Fröhlich, F. and Wibral, M. (2017) Breakdown of local information processing may underlie isoflurane anesthesia effects. PLoS Computational Biology, 13 (6). e1005511. ISSN 1553-734X doi: https://doi.org/10.1371/journal.pcbi.1005511