BDBComp
Parceria:
SBC
A/R+: Um Algoritmo para Inferência em Redes Credais

José Carlos Ferreira da RochaFábio Gagliardi Cozman

This article presents the A/R+ algorithm for approximate inference in credal networks (a framework to represent imprecise probabilistic knowledge). Given a credal network, an inference aims at computing the upper and lower bounds for the marginal probability of each category of a random variable in the network. The proposed algorithm extends the A/R algorithm, an algorithm for approximate inference with interval-valued Bayes nets that can be applied to credal networks. Bounds obtained by A/R+ are significantly more precise than those produced by A/R. Clique no link abaixo para buscar o texto completo deste trabalho na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato: bdbcomp@lbd.dcc.ufmg.br
     Mantida por:
LBD