0000000001264437
AUTHOR
E Wit
Model selection for penalized Gaussian Graphical Models
High-dimensional data refers to the case in which the number of parameters is of one or more order greater than the sample size. Penalized Gaussian graphical models can be used to estimate the conditional independence graph in high-dimensional setting. In this setting, the crucial issue is to select the tuning parameter which regulates the sparsity of the graph. In this paper, we focus on estimating the "best" tuning parameter. We propose to select this tuning parameter by minimizing an information criterion based on the generalized information criterion and to use a stability selection approach in order to obtain a more stable graph. The performance of our method is compared with the state…
Mixing modelling ideas for microarray data
Mixed models have typically been used for modelling structural effects in presence of random variations. These type of models can be used rather naturally when we work with microarray data. In this paper, we shall look at two extensions of the usual mixed effect models.
Cyclic coordinate for penalized Gaussian graphical models with symmetry restriction
In this paper we propose two efficient cyclic coordinate algorithms to estimate structured concentration matrix in penalized Gaussian graphical models. Symmetry restrictions on the concentration matrix are particularly useful to reduce the number of parameters to be estimated and to create specific structured graphs. The penalized Gaussian graphical models are suitable for high-dimensional data.
DgCox: a differential geometric approach for high-dimensional Cox proportional hazard models
Many clinical and epidemiological studies rely on survival modelling to detect clinically relevant factors that affect various event histories. With the introduction of high-throughput technologies in the clinical and even large-scale epidemiological studies, the need for inference tools that are able to deal with fat data-structures, i.e., relatively small number of observations compared to the number of features, is becoming more prominent. This paper will introduce a principled sparse inference methodology for proportional hazards modelling, based on differential geometrical analyses of the high-dimensional likelihood surface.
Robustness of dynamic gene regulatory networks in Neisseria
Gene regulatory networks are made of highly tuned, sparse and dynamical operations. We consider the case of the Neisseria meningitidis bacterium, a causative agent of life-threatening infections such as meningitis, and aim to infer a robust net- work of interactions across sixty proteins based on a detailed time course gene expres- sion study. We consider the problem of estimating a sparse dynamic Gaussian graphical model with L1 penalized maximum likelihood under a structured precision matrix. The structure can consist of specific time dynamics, known presence or absence of links in the graphical model or equality constraints on the parameters. The authors developed a new optimization algo…
Differential geometric LARS via cyclic coordinate descent method
We address the problem of how to compute the coefficient path implicitly defined by the differential geometric LARS (dgLARS) method in a high-dimensional setting. Although the geometrical theory developed to define the dgLARS method does not need of the definition of a penalty function, we show that it is possible to develop a cyclic coordinate descent algorithm to compute the solution curve in a high-dimensional setting. Simulation studies show that the proposed algorithm is significantly faster than the prediction-corrector algorithm originally developed to compute the dgLARS solution curve.