Context-specific independence in graphical log-linear models.
Henrik J. NymanJohan PensarTimo KoskiJukka CoranderPublished in: Comput. Stat. (2016)
Keyphrases
- log linear models
- context specific independence
- probabilistic inference
- probabilistic model
- bayesian networks
- mixture model
- conditional random fields
- domain knowledge
- graphical models
- latent variables
- conditional independencies
- approximate inference
- efficient learning
- message passing
- conditional probability tables
- discriminative training
- max margin
- conditional independence
- random variables
- generative model
- probability distribution
- feature selection
- conditional probabilities
- expectation maximization