Attached is a [tarred] file containing MATLAB *.m files (V4.0) for
demonstrating several ways of parameterizing membership
functions (MF) and learning in fuzzy inference systems (FIS).
Specifically, it includes

	(a) manual tuning of MF's 
	(b) animation of MF's to show the effects of different parameters
	(c) animation of MF's and the overall input/output behavior of FIS.
	(d) learning (parameter identification) in FIS using 
		1. Simplex method
		2. Simplex method + least-squares method

Note that this is only a prototype which is good for education
purpose. Eventually I would like to put the simulation examples in my
papers (especially [1] and [3] listed below) to show how a fuzzy
inference system can be an effective tool for constructing nonlinear
functions from training data, just like neural networks.
(To do so, I need to call C functions inside MATLAB to speed up.)

Related papers
[1] "ANFIS: Adaptive-Network-based Fuzzy Inference Systems", IEEE
    Trans. on Systems, Man and Cybernetics, May 1993.
[2] "Functional Equivalence between Radial Basis Function Networks and
    Fuzzy Inference Systems", Trans. on NN, Jan. 1993.
[3] "Self-Learning Fuzzy Controllers based on Temporal
    Back-Propagation", Trans. on NN, Sept 1992.


=======================================================================
Jyh-Shing Roger Jang			Phone: 510-642-5029
571 Evans, Dept. of EECS		Fax  : 510-642-5775
University of California		Email: jang@eecs.berkeley.edu
Berkeley, CA 94720
=======================================================================


