Bp 210 en
Fuzzy Clustering in Modelling of Nonlinear Systems
Author: Jan Procházka
This paper explores fuzzy clustering methods and their usage for modeling nonlinear systems. We discuss methods fuzzy c-means and Gustafson-Kessel algorithm. Further we focused Takagi-Sugeno models, their principle and assembling using fuzzy clustering methods quoted above. Takagi-Sugeno models were tested on static nonlinear system where we verified aproximation capability and on dynamic discrete nonlinear system, where algorithms were searching for fuzyy clusters in four dimensions.