Grant: Discovery Grants Program - Individual
University of Toronto
$14,000.00
Statistical methods for networks, clusters, and sparse high-dimensional data
EF 2012-13
Résumé
Our research will address the following methodological areas:
1) Constructing network models based on experimental procedures. The procedures we address perturb a variable or variables in a steady-state system and monitor the effect on the system. We have shown that existing differential equation based procedures can be cast as simultaneous equation models from econometric
theory. The work will expand the technology of perturbation experiments, and develop a theory of experimental design for such experiments.
2) Methods relating two or more types of high-dimensional data. We have been working on methods for sparse singular value decomposition, which enable high-dimensional canonical correlation analysis relating two different types of data, such as gene expression and SNP genotype. We will develop these methods, stressing computation, interpretation, and ""oracle properties"" ( abilility to select the correct variables in a sparse model). This will provide exploratory methods for identifying complex relationships between two classes of variables.
3) Prediction based on high-dimensional data. We have been exploring the identification of clusters of variables which are both correlated within cluster and predictive of an outcome variable. We will develop a prediction method incorporating an original supervised clustering algorithm, our current work on filtering variables prior tocluster analysis, and our work on cluster number. This will apply to important applications such as medical diagnosis or prognosis.
4) Cluster analysis for high-dimensional data. We will expoit our current work on sparse singular value decomposition to define clusters based on a collection of sparse eigenvectors. We will extend our methods to two-way clustering, where clusters of variables are defined forsubsets of individuals.
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
121803-2008
Name nom
Tritchler, David
Department département
Public Health, Dalla Lana School of
Organization id
31
Institution établissement
University of Toronto
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2012
Competition year année de concours
2008
Award amount
14000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
Discovery grants programs
Group fr
Programmes de subventions à la découverte
Committee code
14
Committee name en
Statistical Sciences
Committee name fr
Statistique
Area of application code
1205
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Mathematical sciences
Area of application fr
Sciences mathématiques
Research subject code
3005
Research subject group en
Statistics and probability
Research subject group fr
Statistique et probabilité
Research subject en
Biostatistics
Research subject fr
Biostatistiques
Partie
5
Nb partie
5
Application title
Statistical methods for networks, clusters, and sparse high-dimensional data
Keyword
NSERC
Application summary
Our research will address the following methodological areas:
1) Constructing network models based on experimental procedures. The procedures we address perturb a variable or variables in a steady-state system and monitor the effect on the system. We have shown that existing differential equation based procedures can be cast as simultaneous equation models from econometric
theory. The work will expand the technology of perturbation experiments, and develop a theory of experimental design for such experiments.
2) Methods relating two or more types of high-dimensional data. We have been working on methods for sparse singular value decomposition, which enable high-dimensional canonical correlation analysis relating two different types of data, such as gene expression and SNP genotype. We will develop these methods, stressing computation, interpretation, and ""oracle properties"" ( abilility to select the correct variables in a sparse model). This will provide exploratory methods for identifying complex relationships between two classes of variables.
3) Prediction based on high-dimensional data. We have been exploring the identification of clusters of variables which are both correlated within cluster and predictive of an outcome variable. We will develop a prediction method incorporating an original supervised clustering algorithm, our current work on filtering variables prior tocluster analysis, and our work on cluster number. This will apply to important applications such as medical diagnosis or prognosis.
4) Cluster analysis for high-dimensional data. We will expoit our current work on sparse singular value decomposition to define clusters based on a collection of sparse eigenvectors. We will extend our methods to two-way clustering, where clusters of variables are defined forsubsets of individuals.