A study of probabilities and belief functions under conflicting evidence: Comparisons and new methods

  • Mary Deutsch-McLeish
1. Mathematical Theory Of Evidence
Part of the Lecture Notes in Computer Science book series (LNCS, volume 521)


This paper compares the expressions obtained from an analysis of a problem involving conflicting evidence when using Dempster's rule of combination and conditional probabilities. Several results are obtained showing if and when the two methodologies produce the same results. The role played by the normalizing constant is shown to be tied to prior probability of the hypothesis if equality is to occur. This forces further relationships between the conditional probabilities and the prior. Ways of incorporating prior information into the Belief function framework are explored and the results are analyzed. Finally a new method for combining conflicting evidence in a belief function framework is proposed. This method produces results more closely resembling the probabilistic ones.


Belief Functions Probability Theory Conflicting Evidence New Combination Rules 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Charniak, E. (1983) “The Bayesian Basis of Common Sense Medical Diagnosis”. Proc. of the National Conference on Artificial Intelligence, Wash., D.C., pp. 70–73.Google Scholar
  2. Chatolic, P., Dubois, D., Prade, H. (1987). “An Approach to Approximate Reasoning Based on Dempster Rule of Combination”, International Journal of Expert Systems, pp. 67–85.Google Scholar
  3. Dempster, A.P. (1968). “A Generalization of Bayesian Inference”, Journal of the Royal Statistical Society, Series B, 30, pp. 205–247.Google Scholar
  4. Dubois, D., Prade, H. (1988). “Representation and Combination of Uncertainty with Belief Functions and Possibility Measurements”, Computational Intelligence, Vol. 4, pp. 244–264.Google Scholar
  5. Dubois, D., Prade, H. (1985). “Combination and Propagation of Uncertainty with Belief Functions”, IJCAI, 111–113.Google Scholar
  6. Fagin, R., Halpern, J. (1989). “Uncertainty, Belief, and Probability”, IJCAI'89, Proceedings, pp. 1161–1167.Google Scholar
  7. Lindley, D.V. (1985). “Making Decisions”, 2nd Edition, John Wiley.Google Scholar
  8. McLeish, M. (March, 1989). “Extreme Probabilities and Non-Monotonic Reasoning in Probabilistic Logic, Dempster-Shafer and Certainty Factor Models”, submitted, Journal of Approximate Reasoning.Google Scholar
  9. McLeish, M. (1989). “Further Work on Non-monotonicity in the Framework of Probabilistic Logic with an Extension to Dempster-Shafer Theory”, Machine Intelligence and Pattern Recognition, Vol. 8, pp. 23–34.Google Scholar
  10. McLeish, M. (1990). “A Model for Non-Monotonic Reasoning Using Dempster's Rule”, Proceedings of the 6th Uncertainty Management Conference, Cambridge, Mass., pp. 518–528.Google Scholar
  11. Pearl, J. (1988). “Probabilistic Reasoning in Intelligent Systems”, Morgan Kaufmann.Google Scholar
  12. Pearl, J. (1990). “Reasoning with Belief Functions; an Analysis of Compatibility”, Journal of Approximate Reasoning, Vol. 4, pp. 363–389.CrossRefMathSciNetGoogle Scholar
  13. Shafer, G. (1979). “A Mathematical Theory of Evidence”, Princeton University Press, Princeton, NJ.Google Scholar
  14. Zadeh, L.A. (1986). “A Simple View of the Dempster Shafer Theory of Evidence and its Implication for the Rule of Combination”, A.I. Magazine, 7, pp. 85–90.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1991

Authors and Affiliations

  • Mary Deutsch-McLeish
    • 1
  1. 1.Departments of Computing and Information Science/MathematicsUniversity of GuelphGuelphCanada

Personalised recommendations