By Topic

Design for self-organizing fuzzy neural networks using a novel hybrid learning algorithm

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Liang Zhao ; Chinese Acad. of Sci., Beijing ; Fei-Yue Wang

A novel hybrid learning algorithm to implement automatically structure identification and parameter optimization for designing a self-organizing TSK-Type fuzzy neural network (FNN) is proposed in this paper. It includes mean shift clustering algorithm (MSC) and mean firing strength method (MFS) which are employed to identify the network structure of fuzzy neural network (FNN) and the particle swarm optimization enhancing genetic algorithm (PSO-EGA) and the modified back-propagation algorithm (MBP) which are applied to learn the free parameters of it. That is, the MSC is used to partition the input vector space to generate initial network structure. Then the MFS is used to prune the least important rule neurons of initial structure and generate optimal network structure. After the structure identification is completed, the PSO-EGA is adopted to perform a global search in free parameter space of the FNN and seek a near- optimal initial free parameters point for the next stage. Then, it is considered as the initial weights of the FNN and the MBP is used to perform the learning process until a terminal condition is satisfied. The simulation experiment has verified that the proposed hybrid learning algorithm achieves superior performance in learning accuracy than those of some traditional methods.

Published in:

Evolutionary Computation, 2007. CEC 2007. IEEE Congress on

Date of Conference:

25-28 Sept. 2007