Human Expertise, Statistical Models, and Knowledge-Based Systems

  • Dominic A. Clark


The focus of this chapter is the provision of decision support modeled, at least in part, on human expertise. This chapter is divided into sections, each concerned with some aspect of the relationship between human expertise, statistical modeling, and knowledge-based systems, in particular expert systems. The first section provides a comparison of human expert judgment with statistical models, particularly linear regression models, concluding that statistical techniques have been shown to be more accurate than human experts given knowledge of which variables to include in the analysis and the relevant data. The second section provides a brief introduction to expert systems and a comparison of the accuracy of expert systems with human experts. Like statistical models, some expert systems have been shown to be more accurate than human experts. The third section discusses the circumstances under which symbolic representation of knowledge (as rules or frames) or purely statistical approaches are likely to be more useful in building decision aids. Since the utility of a decision support system is a combined function of both the cost of system development and the change in performance of the decision maker(s) using the system, this analysis necessarily raises considerations relating to knowledge engineering and utility of system explanations as well as that of accuracy. The final section summarizes the relative merits and demerits of statistical and symbolic styles of reasoning, emphasizing the complementarity of the two approaches, and provides a brief illustration of how statistical and symbolic reasoning may be combined within a knowledge-based decision support system.


Expert System Human Expert Simple Linear Model Certainty Factor Symbolic Approach 
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.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Armstrong, J. S. (1978). Long range forecasting: From crystal ball to computer. New York: Wiley.Google Scholar
  2. Bowman, E. H. (1963). Consistency and optimality in managerial decision making. Management Science, 9, 310–321.CrossRefGoogle Scholar
  3. Buchanan, B. G. (1986). Expert systems: Working systems and the research literature. Expert Systems, 3(1), 32–51.CrossRefGoogle Scholar
  4. Buchanan, B. G., & Shortliffe, E. H. (1984). Rule-based expert systems: The Mycin experiments of the Stanford Heuristic Programming Project. Reading, Mass.: Addison-Wesley.Google Scholar
  5. Buchanan, B. G., & Smith, R. G. (1988). Fundamentals of expert systems. AnnualReview Computer Science, 3, 23–58.CrossRefGoogle Scholar
  6. Carroll, B. (1987). Expert systems from clinical diagnosis: Are they worth the effort? Behavioral Science, 32, 274–292.PubMedCrossRefGoogle Scholar
  7. Cheeseman, P. (1985). In defense of probability. In Proceedings of International Conference on Artificial Intelligence, Los Angeles, 1002–1009.Google Scholar
  8. Cheeseman, P. (1986). Probabilistic vs. fuzzy reasoning. In Kanal & Lemmer, Uncertainty in artificial intelligence. Amsterdam: Elsevier North Holland.Google Scholar
  9. Cheeseman, P. (1988). An inquiry into computer understanding. Computational Intelligence, 4(1), 58–66.CrossRefGoogle Scholar
  10. Clark, D. A. (1990). Numerical and symbolic approaches to uncertainty management in AI: A review and discussion. Artificial Intelligence Review, 4(2), 109–146.CrossRefGoogle Scholar
  11. Clark, D. A., Fox, J., Glowinski, A. J., & O’Neil, M. (1990). Symbolic reasoning for decision making. In Borcherding, K., Larichev, O., I., & Messick, D. M. (Eds.), Current issues in decision making. Amsterdam: Elsevier.Google Scholar
  12. Dawes, R. M. (1971). A case study of graduate admissions: Application of three principles of human decision making. American Psychologist, 26, 180–188.CrossRefGoogle Scholar
  13. Dawes, R. M. (1979). The robust beauty of improper linear models in decision making. American Psychologist, 34, 571–582.CrossRefGoogle Scholar
  14. Dawes, R. M., & Corrigan, B. (1974). Linear models in decision making. Psychology Bulletin, 81(2), 95–106.CrossRefGoogle Scholar
  15. DeDombal, F. T. (1975). Computer assisted diagnosis of abdominal pain. In Rose, J. & Mitchell, J. (Eds.), Advances in medical computing. New York: Churchill Livingston.Google Scholar
  16. Duda, R. O., & Shortliffe, E. H. (1983). Expert systems research. Science. 220, 261–268.PubMedCrossRefGoogle Scholar
  17. Ellam, S., & Maisey, M. N. (1987). A knowledge-based system to assist in medical image interpretation: Design and evaluation methodology. In Bramer, M. A. (Ed.), Research and development in Expert Systems III. Cambridge University Press.Google Scholar
  18. Fox, J. (1986a). Knowledge, decision making and uncertainty. In Gale, W. A. (Ed.), Artificial intelligence and statistics. Reading, Mass.: Addison-Wesley.Google Scholar
  19. Fox, J. (1986b). Three arguments for extending the framework of probability. In Kanal, L. N. & Lemmer, J. F. (Eds.), Uncertainty and artificial intelligence. Amsterdam: Elsevier B. V.Google Scholar
  20. Fox, J, Barber, D, & Bardhan, K, D. (1980). Alternatives to Bayes. Methods of Information in medicine, 19, 210–215.PubMedGoogle Scholar
  21. Fox, J, O’Neil, M, Glowinski, A., & Clark, D. A. (1989). Decision making as a logical process. In Research and development in expert systems V. Cambridge, University Press.Google Scholar
  22. Fox, J, Clark, D. A, Glowinski, A., & O’Neil, M. (1990). Using predicate logic to integrate qualitative reasoning and classical decision theory. IEEE: Trans Systems, man and cybernetics.Google Scholar
  23. Goldberg, L. R. (1970). Man versus model of man: A rationale plus some evidence for improving clinical inference. Psychological Bulletin, 73(6) 422–432.CrossRefGoogle Scholar
  24. Hammond, K. R. (1987). Toward a unified approach to the study of expert judgment. In Mumpower, J. et al. (Eds.), Expert judgement and expert systems. Heidelberg: Springer-Verlag.Google Scholar
  25. Harmon, P., & King, D. (1985). Expert systems: Artificial intelligence in business. Chichester: Wiley.Google Scholar
  26. Heckerman, D., & Horvitz, E. (1988). The myth of modularity in rule-based systems for reasoning with uncertainty. In Lemmer, I. F. & Kanal, L. N. (Eds.), Uncertainty in Artificial Intelligence 2. Amsterdam: Elsevier North Holland.Google Scholar
  27. Hoffman, P. J. (1960). The paramorphic representation of clinical judgement. Psychological Bulletin, 57(2), 116–131.PubMedCrossRefGoogle Scholar
  28. Kline, P. J., & Dolins, S. B. (1989). Designing expert systems. Chichester: Wiley.Google Scholar
  29. Kramer, B. M. & Mylopoulos, J. (1987). Knowledge representation. In Shapiro, S. C. (Ed.), Encyclopedia of artificial intelligence. Chichester: Wiley.Google Scholar
  30. Lauritzen, S. L., & Spiegelhalter, D. J. (1988). Local computations with probabilities on graphical structures and their application to expert systems (with discussion). Proceedings of the Royal Statistical Society, B, 50, 157–224.Google Scholar
  31. Lenat, D., & Guha, (1990). Building large knowledge based systems. Reading, Mass.: Addison-Wesley.Google Scholar
  32. McDermott, J. (1982). XSEL: A computer sales person’s assistant. Machine Intelligence, 10, 325–337.Google Scholar
  33. Meehl, P. E. (1954). Clinical versus statistical prediction. Minneapolis: University of Minnesota Press.Google Scholar
  34. Meehl, P. E. (1959). A comparison of clinicians with five statistical methods of identifying psychotic MMPI profiles. Journal of Counseling Psychology, 6(2) 102–109.CrossRefGoogle Scholar
  35. Meehl, P. E. (1965). Seer versus sign: The first good example. Journal of Experimental Research in Personality, 1, 27–32.Google Scholar
  36. Minsky, M. (1975). A framework for representing knowledge. In Winston, P. (Ed.), The psychology of computer vision. New York: McGraw-Hill.Google Scholar
  37. Newell, A., & Simon, H. (1972). Human problem solving. Englewood Cliffs, N.J.: Prentice-Hall.Google Scholar
  38. O’Neil, M, Glowinski, A, & Fox, J. (1989). A symbolic theory of decision making applied to several medical tasks. In Lecture Notes in Medical Informatics, 38: Proceedings of the European Conference in AI in Medicine. Berlin: Springer-Verlag.Google Scholar
  39. Patil, R. S, Szolovits, P., & Schwartz, W. B. (1984). Causal understanding of patient illness in medical diagnosis. In Clancey, W. J. & Shortliffe, E. H. (Eds.), Readings in medical artificial intelligence: The first decade. Reading, Mass.: Addison Wesley, pp. 339–360.Google Scholar
  40. Pearl, J (1988). Probabilistic reasoning in intelligent systems. Los Altos, CA: Morgan Kaufmann.Google Scholar
  41. Ramsey, C. L, Reggia, J. A, Nau, D. S., & Ferrentino, A. (1986). A comparative analysis of methods for expert systems. International Journal of Man-Machine Studies, 24, 475–499.CrossRefGoogle Scholar
  42. Rauch-Hindin, W. B. (1986). Artificial intelligence in business, science and industry: Volume II—Applications.” Englewood Cliffs, N.J.: Prentice-Hall.Google Scholar
  43. Reddy, R. (1988). Foundations and grand challenges of artificial intelligence. AI Magazine, Winter, 9–21.Google Scholar
  44. Reggia, J, Tabb, R, Price, T, Banko, M., & Hebel, R. (1984). Computer-aided assessment of transient ischemic attacks. Archives of Neurology, 41, 1248–1254.PubMedGoogle Scholar
  45. Schwartz, S, Griffin, T., & Fox, J. (1989). Clinical expert systems versus linear models. Behavioral Science, 34, 305–311.PubMedCrossRefGoogle Scholar
  46. Shafer, G. (1987). Probability judgement in artificial intelligence and expert systems. Statistical Science, 2 (1), 3–16.Google Scholar
  47. Shanteau, J. (1987). Psychological characteristics of expert decision makers. In Mumpower, J. et al. (Eds.), Expert judgement and expert systems. Heidelberg: Springer-Verlag.Google Scholar
  48. Shortliffe, E. H. (1976). Computer-based medical consultations: Mycin. New York: Elsevier.Google Scholar
  49. Shortliffe, E. H., & Buchanan, B. G. (1975). A model of inexact reasoning in medicine. Mathematical Biosciences, 23, 351–379.CrossRefGoogle Scholar
  50. Smith, S, Fox, M. S., & Ow P. S. (1986). Constructing and maintaining detailed production plans: Investigations into the development of knowledge-based factory scheduling systems. AI Magazine, Fall, 45–61.Google Scholar
  51. Spiegelhalter, D. J. (1986a). A statistical view of uncertainty inexpert systems. In Gale, W. A. (Ed.), Artificial Intelligence and Statistics. Reading, Mass.: Addison-Wesley.Google Scholar
  52. Spiegelhalter, D. J. (1986b). Probabilistic reasoning in predictive expert systems. In Kanal, L. N. & Lemmer, J. F. (Eds.), Uncertainty in artificial intelligence. Amsterdam: North Holland.Google Scholar
  53. Spiegelhalter, D. J., Franklin, R. C. G., & Bull, K. (1989). Assessment, criticism and improvement of imprecise subjective probabilities for a medical expert system. Uncertainty in Artificial Intelligence 5, Elsevier.Google Scholar
  54. Teach, R. L., & Shortliffe, E. H. (1981). An analysis of physician attitudes regarding computer-based clinical consultation systems. Computers in Biomedical Research, 14, 542–558. Reprinted in Buchanan, B. G. & Shortliffe, E. (Eds.), Rule-based expert systems: The Mycin experiments of the Stanford Programming Project. Reading, Mass.: Addison-Wesley.CrossRefGoogle Scholar
  55. Walker, T. C., & Miller, R. K. (1986). Expert systems 1986. Madison, Ga: SEAI Tech. Publ.Google Scholar
  56. Wiggins, N., & Kohen, E. S. (1961). Man vs model of man revisited: The forecasting of graduate school success. Journal of Personality and Social Psychology, 19, 100–106.CrossRefGoogle Scholar
  57. Wyatt, G., & Spiegelhalter, D. (1989). Evaluating medical expert systems: What to test and how? In Talmon, J. and Fox, J. (Eds.), Systems engineering in medicine. Heidelberg: Springer-Verlag.Google Scholar
  58. Wyatt, G., & Spiegelhalter, D. (1990). Evaluating medical expert systems. Proceedings AAAI Spring Symposium on AI in Medicine, 211–215.Google Scholar
  59. Yntema, D. B., & Torgenson, W. S. (1961). Man-computer cooperation in decision requiring common sense. IRE Transactions of the Professional Group on Human Factors in Electronics, 2(1), 20–26.Google Scholar
  60. Yu, V. L., Fagan, M. L., Bennett, S. W., Clancey, W. J., Scott, A. C., Hannigan, J. F., Blum, R. L., Buchanan, B. G., & Cohen, S. N. (1979). An evaluation of Mycin’s advice. Journal of the American Medical Association, 242, 1279–1282. Reprinted in Buchanan, B. G., & Shortliffe, E. H. (Eds.), Rule-based expert systems: The Mycin experiments of the Stanford Heuristic Programming Project. Reading, Mass.: Addison-Wesley.PubMedCrossRefGoogle Scholar

Copyright information

© Plenum Press, New York 1992

Authors and Affiliations

  • Dominic A. Clark
    • 1
  1. 1.Imperial Cancer Research FundAdvanced Computation LaboratoryLondonEngland

Personalised recommendations