ABSTRACT
A graphical belief model uses a graph to capture the structure of the problem. This chapter describes some philosophical issues connected with graphical belief models, and compares uncertainty and imprecision, and subjective and objective views of probability. It explores the role of graphs and independence conditions in graphical models, compares belief function and probability models, and mentions some applications of graphical modelling. After production, the graphical model can be used as a diagnostic expert system to help field maintenance technicians, or to generate likely failure scenarios to train operators. The chapter describes decision making (or artificial intelligence) uses of graphical models, explains the application of graphical models to reliability, and discusses the Belief package for manipulating graphical belief models. In image analysis, the graphical models represent the relationship between pixels and their neighbors and allow a way of removing noise from the image.
