Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Canada Research Chairs

The University of British Columbia

$50,000.00

Statistics of Extremes and Risk

EF 2024-25

Résumé
In many real-life applications, there is a need to have a statistical model able to describe the behaviour of data. This proposal considers situations in which data are multivariate consisting of measurements on several variables, and the interest is in extreme values of these variables that represent risk factors determining stability of the underlying system. Examples of risk factors include losses on financial assets comprising an investment portfolio, or river water levels at multiple locations along a river-bank network subject to potential flooding. The proposed research program encompasses three objectives: (i) construction of flexible multivariate probability models focused on extremes; (ii) development of statistical methodologies on the basis of these models; (iii) application of the proposed methods for risk assessment across several areas including finance and environmental sciences. A starting point for the structure of models will be a class of multivariate probability density functions whose level sets are scaled copies of a shape set characterizing the shape of multivariate data clouds. Such sets are known as homothetic. The shape set determines (tail) dependence properties of the underlying data. The goal is to extend this class of models and associated statistical estimation and inference methods in the following directions: allowing for flexible marginal behaviour of the univariate components, letting the assumption of homotheticity to hold only in the tail region determined by the edge of the data cloud, and considering a novel more general notion of G-homotheticity for a generator matrix G. While theory can be developed for data of any dimension, in practice, more modelling flexibility and computational efficiency can be achieved if a multivariate model is constructed through bivariate blocks. An approach similar to the construction of vine copulas will be investigated. The proposal is motivated by recent advances in statistical approaches that rely on a geometric representation of multivariate extremes through the shape of data clouds. The approach is promising in its ability to break limitations of the classical paradigm to modelling of multivariate extremes, in particular, with regard to the assumptions on the extremal dependence structure among individual variables. This work will lead to improvements in risk assessment due to more flexible and realistic modelling assumptions.
Programme
Canada Research Chairs
Emplacement
British Columbia, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
CRC-2023-00321
Name nom
Nolde, Natalia
Department département
Statistics - The University of British Columbia
Organization id
2
Institution établissement
The University of British Columbia
Province en
British Columbia
Province fr
Colombie-Britannique
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2023
Award amount
50000
Program id
CRC
Program name en
Canada Research Chairs
Program name fr
Chaires de recherche du Canada
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
9998
Committee name en
Canada Research Chair Selection Committee
Committee name fr
Comité de sélection des Chaires de recherche du Canada
Area of application code
9999
Area of application group en
Not available
Area of application group fr
Non disponible
Area of application en
Not available
Area of application fr
Non disponible
Research subject code
9999
Research subject group en
Not available
Research subject group fr
Non disponible
Research subject en
Not available
Research subject fr
Non disponible
Field of research list codes
RDF1010303
Field of research list names en
Statistical theory and modeling
Field of research list names fr
Théorie et modélisation statistiques
Socioeconomic objective list codes
RDS11101
Socioeconomic objective list names en
Finance and economics
Socioeconomic objective list names fr
Finance et économie
Application title
Statistics of Extremes and Risk
Keywords
extreme value theory; statistics of extremes; asymptotic independence; convergence onto a limit set; copulas; time series; quantitative risk management
Application summary
In many real-life applications, there is a need to have a statistical model able to describe the behaviour of data. This proposal considers situations in which data are multivariate consisting of measurements on several variables, and the interest is in extreme values of these variables that represent risk factors determining stability of the underlying system. Examples of risk factors include losses on financial assets comprising an investment portfolio, or river water levels at multiple locations along a river-bank network subject to potential flooding. The proposed research program encompasses three objectives: (i) construction of flexible multivariate probability models focused on extremes; (ii) development of statistical methodologies on the basis of these models; (iii) application of the proposed methods for risk assessment across several areas including finance and environmental sciences. A starting point for the structure of models will be a class of multivariate probability density functions whose level sets are scaled copies of a shape set characterizing the shape of multivariate data clouds. Such sets are known as homothetic. The shape set determines (tail) dependence properties of the underlying data. The goal is to extend this class of models and associated statistical estimation and inference methods in the following directions: allowing for flexible marginal behaviour of the univariate components, letting the assumption of homotheticity to hold only in the tail region determined by the edge of the data cloud, and considering a novel more general notion of G-homotheticity for a generator matrix G. While theory can be developed for data of any dimension, in practice, more modelling flexibility and computational efficiency can be achieved if a multivariate model is constructed through bivariate blocks. An approach similar to the construction of vine copulas will be investigated. The proposal is motivated by recent advances in statistical approaches that rely on a geometric representation of multivariate extremes through the shape of data clouds. The approach is promising in its ability to break limitations of the classical paradigm to modelling of multivariate extremes, in particular, with regard to the assumptions on the extremal dependence structure among individual variables. This work will lead to improvements in risk assessment due to more flexible and realistic modelling assumptions.