Five Calculi of Uncertainty

Terry Bahill
Systems and Industrial Engineering
University of Arizona
Tucson, AZ 85721-0020, USA
terry@sie.arizona.edu
© 1998-2004 Bahill

Decision support systems are different from tradition computer programs because they can reason in the face of incomplete, uncertain, noisy data. Most decision support systems use one of five techniques for dealing with uncertainty: Bayesian statistics, fuzzy set theory, Mycin style certainty factors, neural nets or Dempster-Shafer theory. In this talk we will examine each of these, study their strengths and weakness, and calculate the information and time complexity of each.

Reference: Henkind SJ and Harrison MC, "An analysis of four uncertainty calculi," IEEE Transactions on Systems, Man and Cybernetics, SMC-18:700-714, 1988.
This lecture takes two hours. This lecture is suitable for engineers. This talk requires an overhead projector.