Grant: Discovery Grants Program - Individual
Université Laval
$37,000.00
Multivariate Statistics: Exact inference, copulas and directions
FY 2024-25
Summary
This research project has several parts. This first one is concerned with the simulation of random 0-1 matrices distributed according to the conditional distribution of the dependent 0-1 data matrix given the sufficient statistics for a model for the data. These random matrices are useful to assess the fit of models considered for this data set. Monte Carlo test procedures are obtained by comparing the value of a goodness of statistics calculated on the dependent matrix to simulated values. This procedure applies to investigate the fit of a log-linear model, either to a multiway contingency table or to a capture-recapture data set. They should also be useful to simulate alternatives to a data matrix that protect the confidentiality of the units that are tabulated.
The second part of the project deals with the analysis of angular and spherical data. One objective is to propose a stepwise method to construct multivariate models for data sets featuring angular variables. Methods to assess the dependency between angular variables, independently of their marginal distributions, will be developed. This will be done by constructing semi-parametric estimators for the parameters of the "circula" that characterizes the dependency between the variables. The spherical component of this project is concerned with CT-scanning data on the tree crown of miniature trees. The 3D unit vectors representing the directions of the individual tree branches is the data set for analysis. Models, associated with various growth strategy for the scanned trees, will be constructed and tested.
The third part of the project is concerned with copula models for multivariate data sets. First copula based flexible regression models for hierarchical data will be developed. Copulas will be used to construct cluster specific regression functions and flexible conditional models for the dependent variables, given the explanatory variables and the cluster effect. The definition of constructs using items in a questionnaire is an important research area in the social sciences. We will investigate how copulas can help with the definition of these constructs. New factorial copula models will be investigated for this purpose.
Program
Discovery Grants Program - Individual
Recipient
Location
Quebec, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data
As published
Application id
RGPIN-2023-04740
Name nom
Rivest, Louis-Paul
Department département
Mathématiques et statistique
Organization id
51
Institution établissement
Université Laval
Province en
Québec
Province fr
Québec
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2023
Award amount
37000
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
201
Area of application group en
Natural resources (economic aspects)
Area of application group fr
Ressources naturelles (aspects économiques)
Area of application en
Forestry (silviculture, forest management)
Area of application fr
Foresterie (sylviculture, aménagement forestier)
Research subject code
3000
Research subject group en
Statistics and probability
Research subject group fr
Statistique et probabilité
Research subject en
Statistics and probability
Research subject fr
Statistique et probabilité
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
Multivariate Statistics: Exact inference, copulas and directions
Keywords
capture-recapture models; goodness of fit tests; multivariate models for circular variables; semi-parametric inference; vine-copulas; survey sampling
Application summary
This research project has several parts. This first one is concerned with the simulation of random 0-1 matrices distributed according to the conditional distribution of the dependent 0-1 data matrix given the sufficient statistics for a model for the data. These random matrices are useful to assess the fit of models considered for this data set. Monte Carlo test procedures are obtained by comparing the value of a goodness of statistics calculated on the dependent matrix to simulated values. This procedure applies to investigate the fit of a log-linear model, either to a multiway contingency table or to a capture-recapture data set. They should also be useful to simulate alternatives to a data matrix that protect the confidentiality of the units that are tabulated.
The second part of the project deals with the analysis of angular and spherical data. One objective is to propose a stepwise method to construct multivariate models for data sets featuring angular variables. Methods to assess the dependency between angular variables, independently of their marginal distributions, will be developed. This will be done by constructing semi-parametric estimators for the parameters of the "circula" that characterizes the dependency between the variables. The spherical component of this project is concerned with CT-scanning data on the tree crown of miniature trees. The 3D unit vectors representing the directions of the individual tree branches is the data set for analysis. Models, associated with various growth strategy for the scanned trees, will be constructed and tested.
The third part of the project is concerned with copula models for multivariate data sets. First copula based flexible regression models for hierarchical data will be developed. Copulas will be used to construct cluster specific regression functions and flexible conditional models for the dependent variables, given the explanatory variables and the cluster effect. The definition of constructs using items in a questionnaire is an important research area in the social sciences. We will investigate how copulas can help with the definition of these constructs. New factorial copula models will be investigated for this purpose.