Abstract
Among the many approaches that have been expounded for treating uncertainties in expert systems, certainty factors are very prominent from the standpoint of usage in expert system construction and commercial shell implementation. Although experimental studies indicate that certainty factors can model the behaviors of some experts for some tasks, they also suggest the applicability of certainty factors is by no means universal. Coupling the prominence of the certainty factor approach with empirical evidence that it does not always succeed in modeling human behavior, it is important to explore circumstances under which commonly used certainty combining methods may be more or less beneficial. One such set of circumstances involves magnitude effects. Here, we report on a laboratory experiment that studies certain magnitude effects.
Original language | English |
---|---|
Pages (from-to) | 301-317 |
Number of pages | 17 |
Journal | International Journal of Expert Systems |
Volume | 7 |
Issue number | 3 |
State | Published - 1994 |
ASJC Scopus subject areas
- General Engineering