Modularized Gaussian Graphical Model
Jian Feng Guo, Sijian Wang · 2010
Gaussian graphical models explore the partial correlation structures between random variables and illustrate them using networks. In this paper, we proposed to improve the estimation of Gaussian graphical models by incorporating the unknown modular structure existing in many real world networks. The proposed method simultaneously identifies the modules and estimates the network structure. Several numerical experiments on simulated and real data sets demonstrate that the estimation of the graphical models can be improved by appropriately utilizing prior modular information.