Probabilistic Reasoning in Intelligent Systems

Preface to the Fourth Printing

The upsurge in research in and applications of Bayesian networks has necessitated a large number of refinements within the original pagination of the first printing. Summaries of and references to new developments were added to the bibliographical sections at the end of each chapter and are also reflected in the revised set of exercises. One major development that could not be included is the transition from the probabilistic to the causal interpretation of Bayesian networks, which led to a calculus of interventions (Biometrica 82(4):669–710, December 1995) and counterfactuals (UAI-94, July 1994, 46-54). I hope to introduce these exciting developments in the near future.

J.P.,     Los Angeles, California

February ...