Approximate probability propagation with mixtures of truncated exponentials

Mixtures of truncated exponentials (MTEs) are a powerful alternative to discretisation when working with hybrid Bayesian networks. One of the features of the MTE model is that standard propagation algorithms can be used. However, the complexity of the process is too high and therefore approximate methods, which tradeoff complexity for accuracy, become necessary. In this paper we propose an approximate propagation algorithm for MTE networks which is based on the Penniless propagation method already known for discrete variables. We also consider how to use Markov Chain Monte Carlo to carry out the probability propagation. The performance of the proposed methods is analysed in a series of experiments with random networks.