The Minimax Learning Rate of Normal and Ising Undirected Graphical Models
share
Let G be an undirected graph with m edges and d vertices. We show that d-dimensional Ising models on G can be learned from n i.i.d. samples within expected total variation distance some constant factor of {1, √((m + d)/n)}, and that this rate is optimal. We show that the same rate holds for the class of d-dimensional multivariate normal undirected graphical models with respect to G. We also identify the optimal rate of {1, √(m/n)} for Ising models with no external magnetic field.
READ FULL TEXT