Selecting high-dimensional mixed graphical models using minimal AIC or BIC forests.
David EdwardsGabriel C. G. de AbreuRodrigo LabouriauPublished in: BMC Bioinform. (2010)
Keyphrases
- graphical models
- high dimensional
- model selection
- information criterion
- bayesian information criterion
- maximum likelihood
- model selection criteria
- belief propagation
- models with hidden variables
- cross validation
- probabilistic model
- variable selection
- approximate inference
- probabilistic inference
- markov networks
- random variables
- structure learning
- probabilistic graphical models
- conditional random fields
- bayesian networks
- dimensionality reduction
- regression model
- exact inference
- selection criterion
- statistical inference
- probability model
- conditional independence
- map inference
- hyperparameters
- belief networks
- sample size
- mixture model
- free energy
- high dimensional data
- parameter estimation
- latent variables
- gene expression data
- loopy belief propagation
- undirected graphical models
- pairwise
- feature space