|本期目录/Table of Contents|

[1]杨海燕,文一凭.一种面向特征选择的分类神经网络[J].武汉工程大学学报,2008,(04):114-117.
 YANG Hai yan,WEN Yi ping.A classification neural network oriented to feature selection[J].Journal of Wuhan Institute of Technology,2008,(04):114-117.
点击复制

一种面向特征选择的分类神经网络(/HTML)
分享到:

《武汉工程大学学报》[ISSN:1674-2869/CN:42-1779/TQ]

卷:
期数:
2008年04期
页码:
114-117
栏目:
机电与信息工程
出版日期:
2008-04-30

文章信息/Info

Title:
A classification neural network oriented to feature selection
文章编号:
16742869(2008)04011404
作者:
杨海燕12文一凭3
1.福建工程学院计算机与信息科学系,福建 福州 350014; 2.中南大学信息学院,湖南 长沙 410075;
3.湖南科技大学数学与计算科学学院,湖南 湘潭 411201
Author(s):
YANG Haiyan12WEN Yiping3
1.Department of Computer and Information Science, Fujian University of Technology, Fuzhou 350014,China;
2. College of Information Science and Engineering, Central South University, Changsha 410075,China;
3.School of Mathematics and Computer Science, Hunan University of Science and Technology, Xiangtan 411201,China
关键词:
特征选择分类中心矢量连接权
Keywords:
feature selectionclassificationcenter vectorconnection weight
分类号:
TP 391.41
DOI:
-
文献标志码:
A
摘要:
提出了一种面向特征选择的分类神经网络.该网络对中心矢量和连接权同时学习,其中心矢量是分类的中心,而基于模糊隶属度的权表示特征的重要性,根据权值大小进行特征取舍,这样同时解决了模式识别中的模式分类和特征选择问题.在IRIS数据集上的测试实验表明,该网络能从原始特征中选择重要特征,同时保持最大辨识率.
Abstract:
This paper proposed a new classification neural network. It can learn the center vector and the connect weight simultaneously.The center vector is the center of classification, and the weight based on fuzzy membership represents the importance of feature.The features can be selected according to the final value of the weight, thus solving two major problems in pattern recognition: pattern classification and feature selection. The effectiveness of this method has been validated by IRIS data. The results show that the proposed network can select important features from the original features and maintain quite same performance as using the whole features.

参考文献/References:

[1]何清.模糊聚类分析理论与应用研究进展[A].模糊技术与神经网络选编[C].见刘曾良.北京:北京航空航天大学出版社,1999:126130.
[2]Li R P. Proportional learning vector quantization[J]. Journal of Japan Society for Fuzzy Theory and Systems. 1998,10(6): 11291134.
[3]Anthony M, John R. Fuzzy learning vector quantization for hyperspectral coastal vegetation classification[J]. Remote Sensing of Environment. 2006,(100):512530.
[4]Wu KuoLung,Yang MinShen. A fuzzysoft learning vector quantization[J]. Neurocomputing,2003,(55):681697.
[5]Mitra S, Hayashi Y. Neurofuzzy rule generation:survey in soft computing framework[J]. IEEE Trans Neural Networks,2000,11(3):748768.
[6]贺玲,鲁汉榕.一种改进的模糊学习矢量化神经网络[J].空军雷达学院学报,2001,15(1):3335.
[7]Anderson E. The IRISes of the Gaspe Peninsula[J]. Bull Amer IRIS Soc,1939,(59):25.

相似文献/References:

[1]姜华艳,朱倩.殷墟甲骨文时称的分类及来源[J].武汉工程大学学报,2010,(06):91.[doi:10.3969/j.issn.16742869.2010.06.022]
 JIANG Huayan,ZHU Qian.Classification and sources of temporal in inscriptions of yin[J].Journal of Wuhan Institute of Technology,2010,(04):91.[doi:10.3969/j.issn.16742869.2010.06.022]
[2]张晨跃,刘黎志*,邓开巍,等.基于MapReduce的朴素贝叶斯算法文本分类方法[J].武汉工程大学学报,2021,43(01):102.[doi:10.19843/j.cnki.CN42-1779/TQ.202009022]
 ZHANG Chenyue,LIU Lizhi*,DENG Kaiwei,et al.Text Classification Method of Naive Bayes Algorithm Based on MapReduce[J].Journal of Wuhan Institute of Technology,2021,43(04):102.[doi:10.19843/j.cnki.CN42-1779/TQ.202009022]

备注/Memo

备注/Memo:
收稿日期:20080526
基金项目:福建省自然科学基金 (2006J0017)
作者简介:杨海燕(1980)女,湖南衡山人,助教,博士研究生.
研究方向:图像处理,模式识别,智能计算等.
更新日期/Last Update: