Supporting probability elicitation by sensitivity analysis
When building a Bayesian belief network, generally a huge number of probabilities have to be assessed. We argue that the elicitation of these probabilities can be supported by iteratively performing sensitivity analyses on the network, starting with rough, initial assessments. Giving insight into which probabilities require high accuracy and which do not, performing a sensitivity analysis allows for focusing elicitation efforts on the more critical probabilities of the belief network.
Unable to display preview. Download preview PDF.
- J. Pearl. Probabilistic Reasoning in Intelligent Systems. Networks of Plausible Inference. Morgan Kaufmann, Palo Alto, 1988.Google Scholar
- G. Guida and C. Tasso. Topics in Expert System Design, North-Holland, Amsterdam, 1989.Google Scholar
- D. von Winterfeldt and W. Edwards. Decision Analysis and Behavioral Research. Cambridge University Press, New York, 1986.Google Scholar
- M.G. Morgan and M. Henrion. Uncertainty, a Guide to Dealing with Uncertainty in Quantitative Risk and Policy Analysis. Cambridge University Press, 1990.Google Scholar
- G.F. Cooper. NESTOR: a Computer-based Medical Diagnostic Aid that Integrates Causal and Probabilistic Knowledge. Report HPP-84-48, Stanford University. 1984.Google Scholar
- S.L. Lauritzen and D.J. Spiegelhalter. Local computations with probabilities on graphical structures and their application to expert systems. Journal of the Royal Statistical Society, Series B, vol. 50, 1988, pp. 157–224.Google Scholar
- M.J. Druzdzel and L.C. van der Gaag. Elicitation of probabilities for belief networks: combining qualitative and quantitative information. Eleventh Conference on Uncertainty in Artificial Intelligence, 1995, pp. 141–148.Google Scholar
- V.M.H. Coupé, L.C. van der Gaag, J.D.F. Habbema. Sensitivity analysis: an aid for belief-network quantification, submitted for publication.Google Scholar