< Back to previous page

Project

Asymptotic theory for multidimensional statistics

We study testing model assumptions in high-dimensional regression models without the sparsity assumption that only a few coefficients are important. Combining estimators, predictions or heterogeneous datasets is another topic of focus of which we will study the theoretical properties to arrive at better estimation methods. Several aspects of the domain of functional data will be investigated, such as the use of non and semiparametric estimation methods and change-point analysis. We also investigate variable selection and estimation methods for non-perfect observations, such as censored data from survival analysis, which, for example, can be described by a mixed model to incorporate that part of the population is immune for the event of interest. Also a study of extreme values in a space-time context and of dependencies between multivariate objects via copulas is foreseen. A group of national and international researchers will join forces in this scientific research network. In this way, mathematical statistics provides support to data based decisions in several disciplines, such as (bio)medical and actuarial sciences and bioinformatics.

Date:1 Jan 2017 →  31 Dec 2023
Keywords:Mathematical statistics, high-dimensional data, model testing, model selection, censored data, non- and semiparameteric methods, dependent data, extreme values
Disciplines:Applied mathematics in specific fields, Statistics and numerical methods