logo 150eme Lego EPFL
Votre titre
Ecole Polytechnique Fédérale de Lausanne  Votre sous-titre
        English only      EPFL > SB > IMA > STAT > Mission
Anthony Davison's research group has a wide range of interests in statistical modelling and computational statistics. 
     Some of our work is on statistics of extremes, which are important in applications as varied as climate change, hydrology, material science, and finance.  Particular recent interests have been the development of flexible models for multiple extreme events, the reliable estimation of long-term trends in environmental time series due perhaps to global warming, the modelling of clusters of financial extremes, and comparisons of extreme-value and financial time series approaches to estimation of value at risk.  The statistical tools underlying this work are nonparametric smoothing, Markov chain Monte Carlo methods, bootstrap resampling, and Bayesian and likelihood inference.
     Another part of our work is performed jointly with the NCCR Plant Survival at the University of Neuchatel.  Several members of the group collaborate in various projects associated to this research program, in statistical analysis of microarrays, in modelling of insect choice experiments and of the behaviour of vineyard pests, in tritrophic interactions, and other related topics. 
     Yet further work is in computational aspects of image analysis and signal processing, the main tools being advanced methods of optimization applied to Markov random fields, wavelets, and related topics. 
     We also undertake projects in other areas when the opportunity presents itself.
The group has close links with statoo.com, a statistical consultancy begun by one of our former research students.

©2004 EPFL, 1015 Lausanne, tél. 021 693 25 65, webmaster.stat@epfl.ch
last update: 20-02-2004