Constructing Bayesian network in a changing world

Abstract

Although the process of decision-making has been investigated for centuries, only in the last few decades have investigators systematically addressed how decisions are made in a dynamic, real-world environment One of the most daunting challenges faced by decision support systems is a perpetual change in their environment. Existing decision support methodologies, tools, and frameworks are often difficult to scale up and adapt to changing knowledge. In this paper, a framework that constructs a decision model from knowledge base according to the changes in environment is presented. The generated Bayesian network can be changed as the situation evolves. This paper describes the knowledge representation and model construction process. Bayesian networks combination method is also discussed in details. Copyright © 2005, American Association for Artificial Intelligence (www.aaai.org). All rights reserved.