In recent years, the problem of estimating a sparse inverse covariance matrix in the moderate-to-large dimensional setting has been an important and challenging task in many fields, including genomics, finance and earth sciences. To achieve sparsity in the inverse covariance matrix, methods based on L1 regularization are widely used, but fail to incorporate rich structural information known a priori. In this thesis, we study the problem of sparse inverse covariance estimation in three different settings in which L1 penalization is inappropriate and alternative penalties must be considered. First, we consider the problem of estimating a sparse inverse covariance matrix in the time-ordered data context. L1-penalized likelihood methods penaliz...
In this paper we consider the task of esti-mating the non-zero pattern of the sparse in-verse covari...
This thesis is articulated around two axes. The first one is a contribution to the study of partial ...
Graphical models have established themselves as fundamental tools through which to understand comple...
We propose maximum likelihood estimation for learning Gaussian graphical models with a Gaussian (ℓ[2...
This thesis develops methodology and asymptotic analysis for sparse estimators of the covariance mat...
One of the fundamental tasks in science is to find explainable relationships between observed pheno...
We consider the problem of estimating a sparse dynamic Gaussian graphical model with L1 penalized ma...
International audienceGaussian graphical models are widely utilized to infer and visualize networks ...
We offer a method to estimate a covariance matrix in the special case that both the covariance matri...
Many applications in engineering, sociology, neuroscience, biology, etc. require the use of sophisti...
Spearman's rank correlation is a robust alternative for the standard correlation coefficient. By usi...
The dependency structure of multivariate data can be analyzed using the covariance matrix. In many f...
Gaussian graphical models are of great interest in statistical learning. Because the conditional ind...
This paper studies the sparsistency and rates of convergence for estimating sparse covariance and pr...
Background: Various l(1)-penalised estimation methods such as graphical lasso and CLIME are widely u...
In this paper we consider the task of esti-mating the non-zero pattern of the sparse in-verse covari...
This thesis is articulated around two axes. The first one is a contribution to the study of partial ...
Graphical models have established themselves as fundamental tools through which to understand comple...
We propose maximum likelihood estimation for learning Gaussian graphical models with a Gaussian (ℓ[2...
This thesis develops methodology and asymptotic analysis for sparse estimators of the covariance mat...
One of the fundamental tasks in science is to find explainable relationships between observed pheno...
We consider the problem of estimating a sparse dynamic Gaussian graphical model with L1 penalized ma...
International audienceGaussian graphical models are widely utilized to infer and visualize networks ...
We offer a method to estimate a covariance matrix in the special case that both the covariance matri...
Many applications in engineering, sociology, neuroscience, biology, etc. require the use of sophisti...
Spearman's rank correlation is a robust alternative for the standard correlation coefficient. By usi...
The dependency structure of multivariate data can be analyzed using the covariance matrix. In many f...
Gaussian graphical models are of great interest in statistical learning. Because the conditional ind...
This paper studies the sparsistency and rates of convergence for estimating sparse covariance and pr...
Background: Various l(1)-penalised estimation methods such as graphical lasso and CLIME are widely u...
In this paper we consider the task of esti-mating the non-zero pattern of the sparse in-verse covari...
This thesis is articulated around two axes. The first one is a contribution to the study of partial ...
Graphical models have established themselves as fundamental tools through which to understand comple...