Grant: Discovery Grants Program - Individual
Queen's University
$27,000.00
Statistical Learning and Inference for Sparse and Heterogeneous Functional and Longitudinal Data
EF 2024-25
Résumé
With the advancement of data acquisition technologies, researchers can now collect repeated measurements over a longer period of time much more efficiently. Such data can be found in nearly all areas and are extremely common in fields such as clinical trials, epidemiological studies, public health surveillance and disease modelling. Various research communities analyze repeated measurements differently using longitudinal data analysis (LDA) and functional data analysis (FDA). For longitudinal data, parametric approaches such as GEE and GLMM are commonly used, whereas nonparametric approaches play an analogous role in functional data.
Within FDA, functional regression undoubtably received the most attention in applications and methodological development. Depending on where functional data appear in the regression model, functional regression can be further categorized into scalar-on-function, function-on-scalar and function-on-function regression
It is common to observe responses to interventions in health studies such as clinical trials and epidemiological studies longitudinally. As clinical trials become more patient-centered, quality of life (QoL) has become an imperative endpoint, as a successful intervention should improve both survival and quality of life. Data on quality of life are often collected over a few limited time points during and after treatment. This results in sparse, heterogeneous, and censored data with relatively high missing and/or censoring rates.
Longitudinally observed QoL data are often analyzed using linear mixed models, time-to-deterioration analyses, or simple comparisons at fixed time points. In these types of analyses, the temporal relationship within the data is not fully explored. In addition, the longitudinally observed QoL data are typically treated as predictors, which can be problematic since QoL measures are outcomes of the intervention. This proposal aims to model longitudinal or functional data as outcome variables and other baseline variables (such as treatment information and prognostic biomarkers) as covariates based on longitudinally collected health-related data. A good alternative to address these issues is function-to-scalar regression.
In this proposal, we focus on the following varying-coefficient model with functional response, which has been studied extensively, however, the methods in the literature do not fully account for the commonly observed features in longitudinal health data, such as the sparseness, heterogeneity and censoring. In this proposal, we aim to expand function-on-scalar regression to account for these issues. Specifically, the proposal includes three projects: 1, function-to-scalar regression for sparse data; 2, function-to-scalar regression for heterogeneous data; 3, functional regression for censored functional data.
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-2023-05485
Name nom
Tu, Wei
Department département
Public Health Sciences
Organization id
29
Institution établissement
Queen's University
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2023
Award amount
27000
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
1101
Area of application group en
Health, education and social services
Area of application group fr
Santé publique, éducation et services sociaux
Area of application en
Human health (including medically-related psychological research)
Area of application fr
Santé publique (y compris recherche psychologique liée à la médecine)
Research subject code
3001
Research subject group en
Statistics and probability
Research subject group fr
Statistique et probabilité
Research subject en
Statistical theory
Research subject fr
Théorie de statistique
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
Statistical Learning and Inference for Sparse and Heterogeneous Functional and Longitudinal Data
Keywords
functional data analysis; functional regression; robust statistic; missing data; cancer clinical trials; quality of life; censored data; biostatistics
Application summary
With the advancement of data acquisition technologies, researchers can now collect repeated measurements over a longer period of time much more efficiently. Such data can be found in nearly all areas and are extremely common in fields such as clinical trials, epidemiological studies, public health surveillance and disease modelling. Various research communities analyze repeated measurements differently using longitudinal data analysis (LDA) and functional data analysis (FDA). For longitudinal data, parametric approaches such as GEE and GLMM are commonly used, whereas nonparametric approaches play an analogous role in functional data.
Within FDA, functional regression undoubtably received the most attention in applications and methodological development. Depending on where functional data appear in the regression model, functional regression can be further categorized into scalar-on-function, function-on-scalar and function-on-function regression
It is common to observe responses to interventions in health studies such as clinical trials and epidemiological studies longitudinally. As clinical trials become more patient-centered, quality of life (QoL) has become an imperative endpoint, as a successful intervention should improve both survival and quality of life. Data on quality of life are often collected over a few limited time points during and after treatment. This results in sparse, heterogeneous, and censored data with relatively high missing and/or censoring rates.
Longitudinally observed QoL data are often analyzed using linear mixed models, time-to-deterioration analyses, or simple comparisons at fixed time points. In these types of analyses, the temporal relationship within the data is not fully explored. In addition, the longitudinally observed QoL data are typically treated as predictors, which can be problematic since QoL measures are outcomes of the intervention. This proposal aims to model longitudinal or functional data as outcome variables and other baseline variables (such as treatment information and prognostic biomarkers) as covariates based on longitudinally collected health-related data. A good alternative to address these issues is function-to-scalar regression.
In this proposal, we focus on the following varying-coefficient model with functional response, which has been studied extensively, however, the methods in the literature do not fully account for the commonly observed features in longitudinal health data, such as the sparseness, heterogeneity and censoring. In this proposal, we aim to expand function-on-scalar regression to account for these issues. Specifically, the proposal includes three projects: 1, function-to-scalar regression for sparse data; 2, function-to-scalar regression for heterogeneous data; 3, functional regression for censored functional data.