Design of An EP-Based Neuro-Fuzzy Classification Model

TFIR > Department of Electrical Engineering > conference > Item 987654321/1660

Please use this identifier to cite or link to this item: http://163.15.40.127/ir/handle/987654321/1660

Title:	Design of An EP-Based Neuro-Fuzzy Classification Model
Authors:	Guo, Nai-Ren Kuo, Chao-Lin Tsai, Tzong-Jiy 蔡宗吉 (東方技術學院電機工程系) 郭乃仁
Contributors:	東方技術學院電機工程系
Keywords:	evolutionary programming Fuzzy logic neural networks classification system
Date:	2009-03-26
Issue Date:	2014-04-18 14:40:41 (UTC+8)
Publisher:	Okayama, Japan
Abstract:	A new method for design of a classification system using the feature extraction and evolutionary programming (EP) are discussed. In this paper, a neuro-fuzzy classification model (NFCM) is proposed. The optimal fuzzy membership functions of the NFCM are extracted from the training data using EP. The NFCM contains the feature extraction unit and the inference unit. In order to improve the proposed NFCM, the Weight Revised Algorithm (WRA) is used to regulate the weight grade of the principal output node of the inference unit. The WRA is utilized for generating new weight to be added when additions are required. The performance is also compared to other classifiers tested on the same databases. Computer simulation results demonstrate that the proposed classification model can provide a sufficiently high classification rate in comparison with other models.
Relation:	Proceedings of the 2009 IEEE International Conference on Networking, Sensing and Control, pp.918-922
Appears in Collections:	[Department of Electrical Engineering] conference

Files in This Item:

There are no files associated with this item.

View Licence