Advertisement

Statistical reasoning and learning in knowledge-bases represented as causal networks

  • David J. Spiegelhalter
  • Steffen L. Lauritzen
Conference paper
Part of the Lecture Notes in Medical Informatics book series (LNMED, volume 36)

Summary

A causal network is frequently used as a representation for qualitative medical knowledge, and conditional probability tables on appropriate sets of variables form the quantitative part of the accumulated experience. For fixed probabilities, we describe efficient algorithms for propagating the effects of multiple items of evidence around multiply-connected networks, that allow precise probabilistic revision of beliefs concerning the current patient. As a database accumulates, we require both the quantitative aspects of the model to be updated, as well as to learn about the qualitative structure, and we suggest some formal statistical tools for these problems.

Keywords

Expert System Causal Network Global Comparison Conditional Probability Table Influence Diagram 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Zusammenfassung

Ein ‘kausales Netzwerk’ wird häufig benutzt zur Darstellung von qualitativem Wissen und bedingte Wahrscheinlichkeitstabellen, angewandt auf geeignete Variablengruppen, bilden den quantitativen Anteil der gesammelten Erfahrung. Für feste Wahrscheinlichkeiten beschreiben wir leistungsfähige Algorithmen zur Ausbreitung der Effekte multipler Evidenzitems im Umkreis vielfach verknüpfter Netzwerke, die eine präzise Wahrscheinlichkeitsüberprüfung der Annahmen hinsichtlich der jeweiligen Patienten erlauben. Da sich Daten in einer Datenbank ansammeln, fordern wir sowohl, daß die qualitativen Aspektè des Modells aktualisiert werden, als auch daß das Verständnis für die qualitative Struktur erweitert wird. Weiterhin schlagen wir einige formal statistische Werkzeuge zur Beurteilung dieser Probleme vor.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Andreassen S, Woldbye M, Falck B and Andersen S K (1987) MUNIN–A causal probabilistic network for interpretation of electromyographic findings. In Proceedings of 10th International Joint Conference on Artificial Intelligence, Milan, Morgan Kaufmann, pp 366–372.Google Scholar
  2. Buchanan B G and Shortliffe E H (1984) Rule-based Expert Systems: the MYCIN Experiment of the Stanford Heuristic Programming Project. Reading, Mass: Addison-Wesley.Google Scholar
  3. Cheeseman P (1985) In defense of probability. In Proceedings of 9th International Joint Conference on Artificial Intelligence, Los Angeles, 1002–1009.Google Scholar
  4. Darroch J N, Lauritzen S L and Speed T P ( 1980 Markov fields and log-linear models for contingency tables. Annals of Statistics, 8, 522–539.CrossRefGoogle Scholar
  5. Dempster A P and Almond R G (1988) In discussion of Lauritzen and Spiegelhalter (1988).Google Scholar
  6. Edwards D and Havranek T (1985) A fast procedure for model search in multidimensional contingency tables. Biometrika, 72, 339–351.CrossRefGoogle Scholar
  7. Edwards D and Havranek T (1987) A fast model selection procedure for large families of models. J.Amer.Statist.Assoc., 82, 205–211.Google Scholar
  8. Jensen F V, Andersen S K, Kjaerulff U and Andreassen S (1987) MUNIN: on the case for probabilities in medical expert systems–a practical exercise. In Proceedings of First Conference of European Society for Artificial Intelligence in Medicine; (Fox J, Fieschi M, Engelbrecht R, (eds.)). Heidelberg: Springer-Verlag. pp 149–160.Google Scholar
  9. Jensen F V (1988) In discussion of Lauritzen and Spiegelhalter (1988).Google Scholar
  10. Lauritzen S L and Spiegelhalter D J (1988) Local computation with probabilities on graphical structures, and their application to expert systems (with discussion). J.Roy.Statist.Soc., B, 50 (to appear).Google Scholar
  11. Olesen K G and Andersen S K (1988) In discussion of Lauritzen and Spiegelhalter (1988).Google Scholar
  12. Pearl J (1986a) Fusion, propagation and structuring in belief networks. Artificial Intelligence, 29, 241–288.CrossRefGoogle Scholar
  13. Schwartz W B, Patil R S and Szolovits P (1987) Artificial intelligence in medicine: where do we stand? New England Journal of Medicine, 316, 685–688.PubMedCrossRefGoogle Scholar
  14. Shafer G and Shenoy P (1988) Bayesian and belief function propagation. Working paper No. 192, School of Business, University of Kansas, Lawrence.Google Scholar
  15. Spiegelhalter D J (1986) A statistical view of uncertainty in expert systems. in Artificial Intelligence and Statistics (Gale W (ed.)). Reading, Mass: Addison-Wesley. pp 17–56.Google Scholar
  16. Spiegelhalter D J and Lauritzen S L (1988) Updating probabilities in expert systems. Manuscript.Google Scholar
  17. Tarjan R E and Yannakakis M (1984) Simple linear-time algorithms to test chordality of graphs, test acyclicity of hypergraphs, and selectively reduce acyclic hypergraphs. SIAM J.Comput. 13, 566–579.CrossRefGoogle Scholar
  18. Wermuth N (1976) Model search among multiplicative models Biometrics, 32, 153–163.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1988

Authors and Affiliations

  • David J. Spiegelhalter
    • 1
  • Steffen L. Lauritzen
    • 2
  1. 1.MRC Biostatistics UnitCambridgeEngland
  2. 2.Institute of Electronic SystemsUniversity of AalborgDenmark

Personalised recommendations