Asymptotic Model Selection for Directed Networks with Hidden Variables.
Dan GeigerDavid HeckermanChristopher MeekPublished in: UAI (1996)
Keyphrases
- hidden variables
- model selection
- directed networks
- marginal likelihood
- hyperparameters
- cross validation
- posterior distribution
- probabilistic model
- bayesian networks
- generative model
- exponential family
- latent variables
- em algorithm
- bayesian inference
- communication delays
- parameter estimation
- model selection criteria
- information criterion
- regression model
- gaussian process
- maximum flow
- variable selection
- missing values
- machine learning
- mixture model
- sample size
- feature selection
- incomplete data
- active learning
- bayesian information criterion
- worst case
- data sets