Incorporating Causal Prior Knowledge as Path-Constraints in Bayesian Networks and Maximal Ancestral Graphs.
Giorgos BorboudakisIoannis TsamardinosPublished in: ICML (2012)
Keyphrases
- bayesian networks
- prior knowledge
- causal relationships
- observable variables
- chain graphs
- conditional independencies
- conditional independence
- incorporation of prior knowledge
- inference in bayesian networks
- shortest path
- random variables
- probabilistic reasoning
- causal networks
- causal discovery
- learning bayesian networks
- conditional distributions
- causal independence
- ultimate goal
- independent set
- probability distribution
- directed graph
- conditional probabilities
- bayesian framework
- prior information
- probabilistic model
- labeled data
- generative model
- indirect effects
- bayesian network learning
- constraint graph
- markov blanket
- causal reasoning
- bayesian network classifiers
- linear constraints
- edge weights
- continuous variables
- structure learning
- constraint satisfaction
- graph databases
- weighted graph
- probabilistic inference