Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

University of Toronto

$50,000.00

Computation and Statistical Modelling: A symbiotic relationship

FY 2024-25

Summary
This proposal's  short term aims are i)  methodology development for Bayesian computation and ii) the application of copulas in latent variable models. The long-term objectives are  to  develop MCMC for intractable models and dependence graphs, and to increase the explanatory power of functional data analysis through copula modelling. Statistical inference has benefitted greatly from the design of clever algorithms aimed at either optimizing a complex function for frequentist inference or at sampling from an intractable distribution for Bayesian inference. The vast expansion of data collection capabilities have resulted in unprecedented data volumes being available for statistical analysis which, in turn, have brought new challenges to the computational algorithms that were, until recently, in common use. The first part of the proposal makes use of statistical principles to address the multifarious computational challenges that come from dealing with massive data with dependent structures. The research in this proposal will close a major gap in the existing  toolbox of algorithms for Bayesian statistics and encompasses development of sampling graph structures conditional on  data with complex dependence.  Sampling the discrete space of all graphs is hindered by the lack a a canonical topology and requires innovative ideas in theory and methodology.  It is expected that significant progress in computation will have a marked impact on a number of scientific fields struggling with large volumes of data and complex models. The second part of the proposal will make use of the copulae's flexibility and adaptability to capture complex dependence structures and use them to enrich the class of latent variable models. Traditionally used in situations where the variable of interest cannot be measured directly (e.g., intelligent quotient, climate state, disease status, economy's health, etc.) and one must thus rely on several surrogate variables, the latent variable models have found powerful uses in various fields such as medicine, climate studies, economics or sociology, to name only a few. More recently, focus has shifted on using such models to fuse data that have been collected disparately but concern the same underlying phenomenon. The copula will allow the detection of more subtle effects of the underlying unobserved variable on the interdependencies between the observed ones, thus bringing new understanding on complex phenomena. The completion of these aims will impact  research communities that are active  in Bayesian computation and statistical modelling for and with copulas. The multivalent application of these methods has the potential to enhance statistical learning in multiple areas of exploration that are of interest to Canada.
Program
Discovery Grants Program - Individual
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
RGPIN-2024-04506
Name nom
Craiu, Virgil
Department département
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
50000
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
400
Area of application group en
Environment
Area of application group fr
Environnement
Area of application en
Environment
Area of application fr
Environnement
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
Computation and Statistical Modelling: A symbiotic relationship
Keywords
bayesian inference; markov chain monte carlo; graphical models; dynamic systems; copula models; multiresolution models; composite likelihood; generalized bayes
Application summary
This proposal's  short term aims are i)  methodology development for Bayesian computation and ii) the application of copulas in latent variable models. The long-term objectives are  to  develop MCMC for intractable models and dependence graphs, and to increase the explanatory power of functional data analysis through copula modelling. Statistical inference has benefitted greatly from the design of clever algorithms aimed at either optimizing a complex function for frequentist inference or at sampling from an intractable distribution for Bayesian inference. The vast expansion of data collection capabilities have resulted in unprecedented data volumes being available for statistical analysis which, in turn, have brought new challenges to the computational algorithms that were, until recently, in common use. The first part of the proposal makes use of statistical principles to address the multifarious computational challenges that come from dealing with massive data with dependent structures. The research in this proposal will close a major gap in the existing  toolbox of algorithms for Bayesian statistics and encompasses development of sampling graph structures conditional on  data with complex dependence.  Sampling the discrete space of all graphs is hindered by the lack a a canonical topology and requires innovative ideas in theory and methodology.  It is expected that significant progress in computation will have a marked impact on a number of scientific fields struggling with large volumes of data and complex models. The second part of the proposal will make use of the copulae's flexibility and adaptability to capture complex dependence structures and use them to enrich the class of latent variable models. Traditionally used in situations where the variable of interest cannot be measured directly (e.g., intelligent quotient, climate state, disease status, economy's health, etc.) and one must thus rely on several surrogate variables, the latent variable models have found powerful uses in various fields such as medicine, climate studies, economics or sociology, to name only a few. More recently, focus has shifted on using such models to fuse data that have been collected disparately but concern the same underlying phenomenon. The copula will allow the detection of more subtle effects of the underlying unobserved variable on the interdependencies between the observed ones, thus bringing new understanding on complex phenomena. The completion of these aims will impact  research communities that are active  in Bayesian computation and statistical modelling for and with copulas. The multivalent application of these methods has the potential to enhance statistical learning in multiple areas of exploration that are of interest to Canada.