Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

University of Alberta

$28,000.00

Tensors: new methodology and applications

FY 2024-25

Summary
This program focuses on the development of novel statistical methods for tensor variate data with a particular emphasis on applications to neuroimaging and financial data. Tensors, which are multi-dimensional extensions of matrices, have recently gained increased attention due to their ability to not only represent data structures but can also be used to analyze complex data structures. An increasing number of statisticians are engaged in research on tensors due to the large number of open research questions and it being at the intersection of statistics, mathematics, machine learning, and data science. In this program, we concentrate on specific problems for tensor data. First, we consider creating original statistical methodology and theory for weakly dependent tensor data. This generalization to weakly dependent data is important as it expands the domain of application beyond the independent and identically distributed case (i.i.d.) including all applications with temporal dependency such as meteorological, brain imaging, power grid and bitcoin data. Although the i.i.d. assumption simplifies the theory, it is an incorrect assumption and leads to biases. Second, we propose new change point detection methods for tensor variate data. The new methods utilize a tensor decomposition (to reduce the dimension of the data), and then detects multiple change-points in the second-order (network) structure of the decomposed data using novel segmentation methods. Third, we propose a computationally efficient Bayesian time-varying approach for modeling high-dimensional time series by employing a tensor decomposition for the coefficient matrices at different lags. Fourth, we introduce novel methods that classify subject-specific tensors into groups that correspond to discrete events (such as presence of disease). To summarize, statistical research for tensor data is a rapidly evolving field with a wide range of applications across various research domains. It provides a powerful framework for analyzing complex, multi-dimensional data. The work in this program is fundamental research in statistics, but it also directly impacts other areas such as engineering, economics and the natural sciences, due to the ubiquity of tensors. This increases the potential for impact. A number of the proposed collaborations in this project will ensure that the methodological developments connect to end-users (through software development) and that the results will be of direct practical utility and of importance to society. In particular, as the people of Canada and other countries live longer lives, the number of brain disorders such as Alzheimer's disease is certain to rise dramatically. Resting state functional magnetic resonance imaging (fMRI) is a popular tool to study neurological disorders and the development of our change point/time varying methods will significantly contribute to the understanding of these disorders and the possible identification of biomarkers.
Program
Discovery Grants Program - Individual
Location
Alberta, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
RGPIN-2024-06102
Name nom
Cribben, Ivor
Department département
Department of Accounting and Business Analytics
Organization id
9
Institution établissement
University of Alberta
Province en
Alberta
Province fr
Alberta
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2024
Award amount
28000
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
1210
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Psychology
Area of application fr
Psychologie
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
Tensors: new methodology and applications
Keywords
tensors; change points; weak dependence; networks; time series; longitudinal data; testing; low rank; sparsity; neurostatistics
Application summary
This program focuses on the development of novel statistical methods for tensor variate data with a particular emphasis on applications to neuroimaging and financial data. Tensors, which are multi-dimensional extensions of matrices, have recently gained increased attention due to their ability to not only represent data structures but can also be used to analyze complex data structures. An increasing number of statisticians are engaged in research on tensors due to the large number of open research questions and it being at the intersection of statistics, mathematics, machine learning, and data science. In this program, we concentrate on specific problems for tensor data. First, we consider creating original statistical methodology and theory for weakly dependent tensor data. This generalization to weakly dependent data is important as it expands the domain of application beyond the independent and identically distributed case (i.i.d.) including all applications with temporal dependency such as meteorological, brain imaging, power grid and bitcoin data. Although the i.i.d. assumption simplifies the theory, it is an incorrect assumption and leads to biases. Second, we propose new change point detection methods for tensor variate data. The new methods utilize a tensor decomposition (to reduce the dimension of the data), and then detects multiple change-points in the second-order (network) structure of the decomposed data using novel segmentation methods. Third, we propose a computationally efficient Bayesian time-varying approach for modeling high-dimensional time series by employing a tensor decomposition for the coefficient matrices at different lags. Fourth, we introduce novel methods that classify subject-specific tensors into groups that correspond to discrete events (such as presence of disease). To summarize, statistical research for tensor data is a rapidly evolving field with a wide range of applications across various research domains. It provides a powerful framework for analyzing complex, multi-dimensional data. The work in this program is fundamental research in statistics, but it also directly impacts other areas such as engineering, economics and the natural sciences, due to the ubiquity of tensors. This increases the potential for impact. A number of the proposed collaborations in this project will ensure that the methodological developments connect to end-users (through software development) and that the results will be of direct practical utility and of importance to society. In particular, as the people of Canada and other countries live longer lives, the number of brain disorders such as Alzheimer's disease is certain to rise dramatically. Resting state functional magnetic resonance imaging (fMRI) is a popular tool to study neurological disorders and the development of our change point/time varying methods will significantly contribute to the understanding of these disorders and the possible identification of biomarkers.