Influence Diagrams, Belief Nets and Decision Analysis

These are the proceedings of a conference which reflected the growing interest in the use of influence diagrams, belief nets and graph-related models for prediction and decision analysis in engineering, statistics, operations research, management science, medicine and artificial intelligence. The conference brought together decision theory and operations research analysts, statisticians, computer scientists, engineers and experts from a variety of disciplines, to discuss the theory and applications of influence diagrams (IDs). These people were all from communities that use probabilistic models, mathematical models and structures for the propagation of evidence, statistical inference and prediction. The result is a sample of diverse research results and applications where IDs are being used. IDs were originally developed for use in the practice of decision analysis. When modelling practical problems it is necessary to combine different components of the problem into a coherent and mutually acceptable description.
IDs are a powerful tool to facilitate communication between groups of clients and the decision analyst or modeller and to represent changing assumptions in a graphical manner that reveals difficult independence assumptions. They help to focus on internal dependencies as a whole rather than in disjointed sections. The conference concluded that IDs are valuable for solving difficult problems associated with probabilistic models, representation of large amounts of information, prediction, inference and decision-making.

Author Biography

Robert M. Oliver and James Q. Smith are the authors of Influence Diagrams, Belief Nets and Decision Analysis, published by Wiley.