A Learning Algorithm of Fuzzy Model Based on Improved Fuzzy Clustering and QR Decomposition

Author

Wang, Hongwei ; Gu, Hong

Author_Institution

Sch. of Electron. & Inf. Eng., Dalian Univ. of Technol.

fYear

2006

fDate

24-26 May 2006

Firstpage

1

Lastpage

5

Abstract

In this paper, we proposed a learning algorithm for fuzzy modeling based on the improved fuzzy clustering method and QR decomposition. The improved fuzzy clustering method is confirmed by using a new objective function, which includes the influence on the input variables and the output variables exerting the input space of fuzzy model. Fuzzy inference matrix acquired from improved fuzzy clustering method is analyzed on the basis of QR decomposition of matrix. According to analyzing the redundancy of the matrix, the structure of fuzzy system is confirmed in the paper. The structure and parameters of fuzzy model are estimated by means of the proposed algorithm. We demonstrate the performance of the proposed algorithm by using the simulating result of the nonlinear system

Keywords

fuzzy set theory; inference mechanisms; learning (artificial intelligence); matrix algebra; nonlinear systems; QR decomposition; fuzzy clustering methods; fuzzy inference matrix; fuzzy model; learning algorithm; nonlinear system; objective function; Clustering algorithms; Clustering methods; Fuzzy control; Fuzzy systems; Inference algorithms; Input variables; Matrix decomposition; Nonlinear systems; Parameter estimation; Takagi-Sugeno model;

fLanguage

English

Publisher

ieee

Conference_Titel

Industrial Electronics and Applications, 2006 1ST IEEE Conference on

Conference_Location

Singapore

Print_ISBN

0-7803-9513-1

Electronic_ISBN

0-7803-9514-X

Type

conf

DOI

10.1109/ICIEA.2006.257094

Filename

4025711