Contemporary Artificial Intelligence

Chapter 1 2

Learning Probabilistic

M odel Param eters

Until the early 1990s, the DAG in a Bayesian net work was ordinarily hand-

constructed by a domain expert. Then the conditional probabilities were as-

sessed b y the expert, learned from data, or obtained using a combination of

both techniques. Eliciting Bayesian networks from experts can be a laborious

and difficult process in the case of large networks. As a result, researchers

developed methods that could learn the DAG from data. Furthermore, they

formalized methods for learning the conditional probabilities from data. In a

Bayesian network, the conditional probability distributions are called the ...