Probabilistic Inference in Artificial Intelligence: The Method of Bayesian Networks

摘要: Bayesian networks are formalisms which associate a graphical representation of causal relationships and an associated probabilistic model. They allow to specify easily consistent model from set local conditional probabilities. In order infer the probabilities some facts, given observations, inference algorithms have be used, since size models is usually large. Several such methods described illustrated. Less advanced related problems, namely learning, validation, continuous variables, time, briefly discussed. Finally, between field other scientific domains reviewed.