Building Models

Abstract

The framework of Bayesian networks is a very efficient language for building models of domains with inherent uncertainty. However, as can be seen from the calculations in Section 2.6, it is a tedious job to perform evidence transmission even for very simple Bayesian networks. Fortunately, software tools that can do the calculation job for us are available. In the rest of this book, we assume that the reader has access to such a system (some URLs are given in the Preface). Therefore, we can start by concentrating on how to use Bayesian networks in model building and defer a presentation of methods for probability updating to Chapter 4.