Biological Network inference by quantitative and nonparametric modeling
Florence d’Alché-Buc
19 November 2015, 00h00 Salle/Bat : 465/PCRI-N
Contact :
Activités de recherche : Biologie des systèmes
Résumé :
We consider the task of biological network inference and re-visit this task using the idea of Jacobian estimation. We propose a new nonparametric approach to vector autoregressive models that imposes sparsity on the elements of the Jacobian. The method is based on the definition of multiple output regression models and new algorithms based on proximal gradients. We show how it applies to gene regulatory network inference and beyond, this application to any problem where some dependency relationships between state variables are searched.
