A REVIEW ON BAYESIAN STRUCTURE LEARNING IN GAUSSIAN GRAPHICAL MODELS

An accurate understanding of complicated relations among numerous variables isof significant importance in science. One attractive procedure to this task is Gaussian graphicalmodels (GGMs), which lately many improvements have been carried out on it. GGMs describethe conditional independence among variables by means of the presence or absence of edges inthe related graph. In this paper, we recap a Bayesian method for structure learning of GGMsbased on the Birth-Death MCMC (BDMCMC) algorithm. We show the application of thismethod on a simulated dataset.

کلیدواژه ها:

Bayesian structure learning ، Gaussian graphical models ، Birth-Death Markov chain Monte Carlo