Methods for the Evaluation of Expert Systems
Expert systems have their origins in traditional data processing, which is based on the definition of the appropriate representation for the sequence of operations and for the data. In conventional systems, the classification of sets of similar entities is usually achieved by providing a variable name and indexing. Arrays are normally used to define relationships. In a conventional control structure what heppens next is predefined by the program. The knowledge processing of a human expert can be presented in different forms, because traditional representation techniques are inadequate to cope with the problems arising in such systems. One approach to representing classes and relations can be achieved, for example, through the use of a predicate calculus and the control structure can be the production system. The productions are the set of rules and the control structure determines what rule is tried next. If the control structure is domain-independent, it could be very simple. In many problems it is convenient to allow a rule to be considered only in certain circumstances. The control structure of such a system is more complex and conventional algorithms are not well suited because in real situations they have a long response time.
Unable to display preview. Download preview PDF.
- Davis, R., 1982, “Expert Systems: Where Are We and Where Do We Go From Here?” A.I. Memorandum 665, Artificial Intelligence Laboratory, MIT.Google Scholar
- Masaryk, P., Pig, P., Urbánek, T., and Tkácik, J., 1987, The consulting system in rheumatology, in: “13th SCS Symposium, PieStany, Czechoslovakia”: 107.Google Scholar
- Peterson, J. L., 1981, “Petri Nets Theory and the Modeling of Systems”, Prentice-Hall, Englewood Cliffs, NJ.Google Scholar
- Píš, P., Mesiar, R., and Horâk, P., 1985, “Expert Systems’ Evaluation”, Technical Report 04–70–85, SF SVST, gUNZ, Bratislava (in Slovak).Google Scholar
- Píš, P., Horák, P., and Mesiar, R., 1986, “Expert Systems and Petri Nets”, Technical Report 04–50–86, SF SVST, SUNZ, Bratislava (in Slovak).Google Scholar
- Píš, P., and Mesiar, R., 1987a, Fuzzy model of inexact reasoning in medical consulting systems, in: “Proc. 1st Joint IFSA-EC and Euro-WG Workshop on Progress in Fuzzy Sets in Europe, Warsaw, 1986”.Google Scholar