Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

Simon Fraser University

$59,000.00

Interpretable, Flexible and Scalable Methods for Functional Data Analysis

FY 2023-24

Summary
Functional data analysis (FDA) is a growing statistical field for analyzing curves, images, or any multidimensional functions, in which each random function is treated as a sample element. Functional data is found commonly in many applications such as time-course gene expression data and brain imaging data. The long-term goal of my research program is to develop interpretable, flexible and scalable FDA methods for high-dimensional functional and scalar variables. I will establish foundational frameworks, design effective regularization schemes and develop computationally efficient methods with user-friendly software for these FDA methods. My first short-term objective is to propose representation learning methods for functional data. We will develop functional nonlinear learning (FunNoL) methods to sufficiently represent multivariate functional data in a low-dimensional feature space by using neural networks. Furthermore, I will merge a classification model for enriching the ability of representations in predicting curve labels. The proposed FunNoL method will achieve better classifications than FPCA, especially in the multivariate functional data setting. My second short-term objective is to develop dynamic survival prediction using longitudinal images. This objective is motivated by predicting the progression of Alzheimer's disease (AD) based on a series of longitudinally observed brain scan images. We will develop a FPCA method to estimate the multidimensional functional data by estimating FPCs and the associated FPC scores in a regression framework. The FPCs scores will be used as features for dynamic survival prediction. My third short-term objective is to estimate derivatives from sparse functional data. Many applications require to estimate the derivatives of underlying trajectories such as longitudinal growth study and online auction data analysis. We will develop methods to directly estimate the optimal empirical basis functions for the derivative functions in the best approximation framework rather than eigen-decomposing the covariance function. My last short-term objective is to develop interpretable functional regression models. Our method will detect the local regions in which the functional covariates have significant effect on quantiles and estimate the functional quantile regression model simultaneously. This method will be applied to study how climate change and other factors affect the crop yields. The proposed research is likely to impact the field of statistics through enabling groundbreaking advances in FDA. Many computationally efficient methods and interpretable models would be developed to analyze large-scale functional data. We will also publish user-friendly software for our new methods, which is available to the general public and will have a broad impact in research and industry. The proposed research provides a wide variety of cutting-edge thesis topics for HQP training.
Program
Discovery Grants Program - Individual
Location
British Columbia, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
RGPIN-2023-04057
Name nom
Cao, Jiguo
Department département
Department of Statistics and Actuarial Science,
Organization id
5
Institution établissement
Simon Fraser University
Province en
British Columbia
Province fr
Colombie-Britannique
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2023
Competition year année de concours
2023
Award amount
59000
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
1201
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Life sciences (including biotechnology)
Area of application fr
Sciences de la vie (y compris biotechnologie)
Research subject code
9999
Research subject group en
Not available
Research subject group fr
Not available
Research subject en
Non disponible
Research subject fr
Non disponible
Application title
Interpretable, Flexible and Scalable Methods for Functional Data Analysis
Keyword
Functional data analysis; Functional principal component analysis; Functional quantile regression
Application summary
Functional data analysis (FDA) is a growing statistical field for analyzing curves, images, or any multidimensional functions, in which each random function is treated as a sample element. Functional data is found commonly in many applications such as time-course gene expression data and brain imaging data. The long-term goal of my research program is to develop interpretable, flexible and scalable FDA methods for high-dimensional functional and scalar variables. I will establish foundational frameworks, design effective regularization schemes and develop computationally efficient methods with user-friendly software for these FDA methods. My first short-term objective is to propose representation learning methods for functional data. We will develop functional nonlinear learning (FunNoL) methods to sufficiently represent multivariate functional data in a low-dimensional feature space by using neural networks. Furthermore, I will merge a classification model for enriching the ability of representations in predicting curve labels. The proposed FunNoL method will achieve better classifications than FPCA, especially in the multivariate functional data setting. My second short-term objective is to develop dynamic survival prediction using longitudinal images. This objective is motivated by predicting the progression of Alzheimer's disease (AD) based on a series of longitudinally observed brain scan images. We will develop a FPCA method to estimate the multidimensional functional data by estimating FPCs and the associated FPC scores in a regression framework. The FPCs scores will be used as features for dynamic survival prediction. My third short-term objective is to estimate derivatives from sparse functional data. Many applications require to estimate the derivatives of underlying trajectories such as longitudinal growth study and online auction data analysis. We will develop methods to directly estimate the optimal empirical basis functions for the derivative functions in the best approximation framework rather than eigen-decomposing the covariance function. My last short-term objective is to develop interpretable functional regression models. Our method will detect the local regions in which the functional covariates have significant effect on quantiles and estimate the functional quantile regression model simultaneously. This method will be applied to study how climate change and other factors affect the crop yields. The proposed research is likely to impact the field of statistics through enabling groundbreaking advances in FDA. Many computationally efficient methods and interpretable models would be developed to analyze large-scale functional data. We will also publish user-friendly software for our new methods, which is available to the general public and will have a broad impact in research and industry. The proposed research provides a wide variety of cutting-edge thesis topics for HQP training.
Simon Fraser University - $59,000 Grant (2023-24) | Canada Spends