Integração de evidências em redes credais e a regra de Jeffrey

Translated title of the contribution: Evidence Integration in Credal Networks and Jeffrey's Rule

Jose Carlos Ferreira da Rocha, A.M. Guimarães, C.P. de Campos

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Credal networks provide a scheme for dealing with imprecise probabilistic models. The inference algorithms often used in credal networks compute the interval of the posterior probability of an event of interest given evidence of the specific kind -- evidence that describe the current state of a set of variables. These algorithms do not perform evidential reasoning in case of the evidence must be processed according to the conditioning rule proposed by RC Jeffrey. This paper describes a procedure to integrate evidence with Jeffrey's rule when performing inferences with credal nets.
Original languagePortuguese
Pages (from-to)251-265
Number of pages15
JournalRevista de Informática Teórica e Aplicada
Volume18
Issue number2
DOIs
Publication statusPublished - 2011
Externally publishedYes

Bibliographical note

(in portuguese)

Fingerprint Dive into the research topics of 'Evidence Integration in Credal Networks and Jeffrey's Rule'. Together they form a unique fingerprint.

  • Cite this