A global approach for learning sparse Ising models

Daniela De Canditiis

We consider the problem of learning the link parameters as well as the structure of a binary-valued pairwise Markov model. Under sparsity assumption, we propose a method based on $…