Re: Objective Meaning of Fuzzy sets

Amendment9 (amendment9@aol.com)
Tue, 5 Mar 1996 19:54:20 +0100


Wouldn't the probability distribution that you choose also be based upon
certain assumptions and beliefs? Would you choose the same probability
distribution for disjoint sample spaces. For instance a Pigmy culture
versus an Austrian? The sample space itself is a product of beliefs.
Isn't it the case, that Reality - what is observed - is transformed by
the observer?
Right now, I'm somewhat satisfied with the possibility of working with
probability spaces at the corners of a hypercube.

But what your driving at is relevant: a sort of meta-rule construct for
fuzzy expert systems. Such meta-rules may help determine the
degree of confidence that the system has in giving its advice or
control rule.

Ellen