In this paper, we propose the plausibility transformation method for translating Dempster-Shafer (D-S) belief function models to probability models, and describe some of its properties. There are many other transformation methods used in the literature for translating belief function models to probability models. We argue that the plausibility transformation method produces probability models that are consistent with D-S semantics of belief function models, and that, in some examples, the pignistic transformation method produces results that appear to be inconsistent with Dempster's rule of combination.

The goal of this paper is to compare the similarities and differences between Bayesian and belief function reasoning. Our main conclusion is that although there are obvious differences in semantics, representations, the rules for combining and marginalizing representations, there are many similarities. We claim that the two calculi have roughly the same expressive power. Each calculus has its own semantics that allow us to construct models suited for these semantics. Once we have a model in either calculus, one can transform it to the other by means of a suitable transformation.

An important class of hybrid Bayesian networks are those that have conditionally deterministic variables (a variable that is a deterministic function of its parents). In this case, if some of the parents are continuous, then the joint density function does not exist. Conditional linear Gaussian (CLG) distributions can handle such cases when the deterministic function is linear and continuous variables are normally distributed. In this paper, we develop operations required for performing inference with conditionally deterministic variables using relationships derived from joint cumulative distribution functions (CDF's). These methods allow inference in networks with deterministic variables where continuous variables are non-Gaussian.

Abstract. Recently, we proposed a new method called the plausibility transformation method to convert a belief function model to an equivalent probability model. In this paper, we compare the plausibility transformation method with the pignistic transformation method. The two transformation methods yield qualitatively different probability models. We argue that the plausibility transformation method is the correct method for translating a belief function model to an equivalent probability model that maintains belief function semantics.

Mixtures of truncated exponentials (MTE) potentials are an alternative to discretization and Monte Carlo methods for solving hybrid Bayesian networks. Any probability density function (PDF) can be approximated by an MTE potential, which can always be marginalized in closed form. This allows propagation to be done exactly using the Shenoy-Shafer architecture for computing marginals, with no restrictions on the construction of a join tree. This paper presents MTE potentials that approximate standard PDF's and applications of these potentials for solving inference problems in hybrid Bayesian networks. These approximations will extend the types of inference problems that can be modeled with Bayesian networks, as demonstrated using three examples.