Optimized FCM-based radial basis function neural networks: A comparative analysis of LSE and WLSE method

Wook Dong Kim*, Sung Kwun Oh, Wei Huang

*此作品的通讯作者

科研成果: 书/报告/会议事项章节会议稿件同行评审

2 引用 (Scopus)

摘要

In this paper, we introduce a new architecture of optimized FCM-based Radial Basis function Neural Network by using space search algorithm and discuss its comprehensive design methodology. As the consequent part of rules of the FCM-based RBFNN model, four types of polynomials are considered. The performance of the FCM-based RBFNN model is affected by some parameters such as the number of cluster and the fuzzification coefficient of the fuzzy clustering (FCM) and the order of polynomial standing in the consequent part of rules, we are required to carry out parametric optimization of network. The space evolutionary algorithm(SEA) being applied to each receptive fields of FCM-based RBFNN leads to the selection of preferred receptive fields with specific local characteristics available within the FCM-based RBFNN. The performance of the proposed model and the comparative analysis between WLSE and LSE are illustrated with by using two kinds of representative numerical dataset.

源语言英语
主期刊名Advances in Neural Networks - ISNN 2010 - 7th International Symposium on Neural Networks, ISNN 2010, Proceedings
207-214
页数8
版本PART 1
DOI
出版状态已出版 - 2010
已对外发布
活动7th International Symposium on Neural Networks, ISNN 2010 - Shanghai, 中国
期限: 6 6月 20109 6月 2010

出版系列

姓名Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
编号PART 1
6063 LNCS
ISSN(印刷版)0302-9743
ISSN(电子版)1611-3349

会议

会议7th International Symposium on Neural Networks, ISNN 2010
国家/地区中国
Shanghai
时期6/06/109/06/10

指纹

探究 'Optimized FCM-based radial basis function neural networks: A comparative analysis of LSE and WLSE method' 的科研主题。它们共同构成独一无二的指纹。

引用此