Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Grants Program - Individual

University of Toronto

$55,000.00

Estimation and inference in extremes of high-dimensional and complex data

EF 2024-25

Résumé
Extreme value theory seeks to build mathematically founded and practically relevant models for extreme events. Univariate extreme value theory describes the tails of a single random variable and is well studied. The multivariate case still presents substantial challenges, this is especially true for extremes in high dimensions. Broadly speaking, this proposal will focus on exploring various modeling, estimation and inference procedures for extremes in the high-dimensional regime where the number of available data can be moderate or small relative to the dimension of the problem. Understanding the simultaneous extreme behavior of several quantities is of crucial importance to many problems that society is facing today, ranging from risk management and resilience to financial crises (where simultaneous large losses are a major concern and realistic models for the dependence between extreme losses are need) to flood risk management and modeling of extreme temperatures/precipitation due to climate change. There are two classical approaches to extreme-value modeling: the peak over threshold (POT) and the block maxima (BM) approach. Which of the two approaches is better suited for the problem at hand depends on the data collection mechanism. This proposal will focus on developing theoretically grounded methodology for modeling and estimation of multivariate dependence structures in the POT and BM framework under sparsity assumptions. While sparse and interpretable models form a ubiquitous toolbox in areas ranging from Finance to Biology and Environmental Sciences, the developments of such tools in extreme value theory is still in its infancy. The proposed research agenda aims to advance our fundamental understanding of such models for extremes and provide practitioners with reliable inference tools which form the backbone of any sound statistical analysis. A particular focus will lie on graphical models which provide sparse and interpretable methodology by linking a graphical structure with conditional independence relationships (for undirected graphs) or causal effects (for directed graphs). The work in this proposal will equip applied researchers in areas ranging from Finance to Hydrology with tools to gain a better understanding of extremes in high dimensions. This will eventually help to reduce the risk of extreme events in setting where interventions are possible and better prepare for dealing with extreme events where such events are unavoidable.
Programme
Discovery Grants Program - Individual
Bénéficiaire
Emplacement
Ontario, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
RGPIN-2024-05528
Name nom
Volgushev, Stanislav
Department département
Department of Statistical Sciences
Organization id
31
Institution établissement
University of Toronto
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2024
Award amount
55000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1508
Committee name en
Mathematics and Statistics
Committee name fr
Mathématiques et statistique
Area of application code
1200
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Advancement of knowledge
Area of application fr
Avancement des connaissances
Research subject code
3011
Research subject group en
Statistics and probability
Research subject group fr
Statistique et probabilité
Research subject en
Multivariate analysis
Research subject fr
Analyse multivariée
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Estimation and inference in extremes of high-dimensional and complex data
Keywords
multivariate extreme value theory; graphical modeling; structure learning; peak over threshold methods; block maxima methods; high dimensional statistics; dependence modeling
Application summary
Extreme value theory seeks to build mathematically founded and practically relevant models for extreme events. Univariate extreme value theory describes the tails of a single random variable and is well studied. The multivariate case still presents substantial challenges, this is especially true for extremes in high dimensions. Broadly speaking, this proposal will focus on exploring various modeling, estimation and inference procedures for extremes in the high-dimensional regime where the number of available data can be moderate or small relative to the dimension of the problem. Understanding the simultaneous extreme behavior of several quantities is of crucial importance to many problems that society is facing today, ranging from risk management and resilience to financial crises (where simultaneous large losses are a major concern and realistic models for the dependence between extreme losses are need) to flood risk management and modeling of extreme temperatures/precipitation due to climate change. There are two classical approaches to extreme-value modeling: the peak over threshold (POT) and the block maxima (BM) approach. Which of the two approaches is better suited for the problem at hand depends on the data collection mechanism. This proposal will focus on developing theoretically grounded methodology for modeling and estimation of multivariate dependence structures in the POT and BM framework under sparsity assumptions. While sparse and interpretable models form a ubiquitous toolbox in areas ranging from Finance to Biology and Environmental Sciences, the developments of such tools in extreme value theory is still in its infancy. The proposed research agenda aims to advance our fundamental understanding of such models for extremes and provide practitioners with reliable inference tools which form the backbone of any sound statistical analysis. A particular focus will lie on graphical models which provide sparse and interpretable methodology by linking a graphical structure with conditional independence relationships (for undirected graphs) or causal effects (for directed graphs). The work in this proposal will equip applied researchers in areas ranging from Finance to Hydrology with tools to gain a better understanding of extremes in high dimensions. This will eventually help to reduce the risk of extreme events in setting where interventions are possible and better prepare for dealing with extreme events where such events are unavoidable.