Programme leaders: S.A. van de Geer, A.W. van der Vaart
This program studies the statistical modelling and analysis of stochastic phenomena. Its main focus point is the development of nonparametric models for complex situations, where data may be high-dimensional or with little structure, arise from observational studies rather then from experimentally controlled studies, and be subject to censoring, missing, or partially observed. Such data may come for instance from epidemiological and medical studies, life sciences, economic studies, satellites, financial markets or industrial settings.
The programme explores various new procedures and infinite-dimensional models. The classical statistical theory needs to be extended, and sometimes in fact abandoned because not being appropriate in the context of large data sets. Concepts such as optimality and efficiency are given a new dress, and parsimony is no longer considered as essential. The study of computational aspects and algorithms is also an important topic within the programme. The rapid increase of computer power makes complex modelling and computation possible, but also quite urgent due to the enormous amounts of data that are collected. Apart from the general mathematical statistical and computational theory, the programme also investigates practical implementations in numerous concrete settings.
Some keywords are: classification and statistical learning, computer simulation, coarsening at random, curve estimation, efficiency bounds, empirical processes, experimental design, fair estimators, financial models. inverse problems, likelihood and Bayes methods, longest increasing subsequence and Hammersley's process, model selection and adaptive methods, multivariate robust methods, non- and semiparametric models, resampling techniques, statistical quality control, statistics of extremes, stochastic geometry, time series and dependent processes.
Practical work carried out in this programme includes: demography, DNA microarrays, finance, growth of cancer cells, imaging, ion channels, statistical process control in industry, the spread of aids, the spread of herbs, and tomography.
The research group has numerous international contacts and an excellent international reputation. The reseachers form a notable group within the statistical community, which is a mix of mathematicians, theoretical and applied statisticians and data-analysts. The programme organizes regular international workshop with help of the Stieltjes Institute. At a national level, the research group has close interactions with other projects, in particular probability, and has good connections with the statisticians in the Mathematical Research Institute, and with the Vereniging voor Statistiek. There is a strong involvement with the Aio-network stochastics, which organizes the national graduate program in stochastics. The European research institute Eurandom in Eindhoven is an important meeting point for the researchers in this programme.