Training a FCM?

George Kalamaras (gsk1@psu.edu)
Tue, 27 Feb 1996 18:53:38 +0100


I plan to use FCM for developing a construction specific
classifcation system for rock engineering. Initially I
thought using an expert system, but when I came across
the very attractive concept of FCM I decided to pursue
the fuzzy representation of qualitative unceratinty instead
of probabilities and expert system. The problem is that
I am a novice in this area, and my literature survey has
not revealed so far how a FCM can be trained. I was wondering if
someone can direct me to a reference, other than the Kosko's
Neural Networks and Fuzzy Systems (which is beyond my understanding).
I looked to old issues of AI Expert, but I was able to find
more qualitative information rather than an arithmetic exercise.
I would appreciate any help.

Regards, George S. Kalamaras