|
|||||
|
|
| 基于模糊预处理的聚类神经网络模型设计 | |
| 引用本文： | 原思聪,刘道华,刘宏兵,王发展. 基于模糊预处理的聚类神经网络模型设计[J]. 西安建筑科技大学学报(自然科学版), 2005, 37(2): 274-277 |
| 作者姓名： | 原思聪 刘道华 刘宏兵 王发展 |
| 作者单位： | 1. 西安建筑科技大学机电学院,陕西,西安,710055 2. 西安建筑科技大学机电学院,陕西,西安,710055;信阳师范学院计算机科学系,河南,信阳,464000 3. 信阳师范学院计算机科学系,河南,信阳,464000 |
| 基金项目： | 陕西省教育厅专项基金资助项目 (0 3 JK117)，陕西省教育厅产业化基金资助项目 (0 4JC2 1) |
| 摘 要： | 在传统的ART网络及BP网络基础上分别作了改进，有机地结合了两种网络的优越性，并对输入数据进行模糊预处理，从而构建了新的比较健壮的网络模型，并具体给出了网络的模型结构及工作原理．该网络的特点是：检测故障的类别数可以增加，能累计增加新的样本类节点，且能分阶段训练，对故障的分类效果好． |
| 关 键 词： | ART网络 BP网络 模糊预处理 聚类 分时训练 |
| 文章编号： | 1006-7930(2005)02-0274-04 |
| 修稿时间： | 2004-08-14 |
The design of clustering neural network based on fuzzy pretreatment |
|
| YUAN Si-Cong,LIU Dao-hua,LIU Hong-bing,WANG Fa-Zhan. The design of clustering neural network based on fuzzy pretreatment[J]. Journal of Xi'an University of Architecture & Technology, 2005, 37(2): 274-277 | |
| Authors: | YUAN Si-Cong LIU Dao-hua LIU Hong-bing WANG Fa-Zhan |
| Affiliation: | YUAN Si-cong~ |
| Abstract: | This paper presented improvement on the back-propagation network (BP) on the basis of traditional adaptive resonance (ART), so that the superiority of the two network is integrated, and had fuzzy pretreatment to the input data is acquired,building up new stable network model, and giving the structure and the working principle of the network model. The network is characterisfic of increasing the types of the inspecting obstacle properly, accumulating the new knot of sample species, and training the network for different phase. This network is good at classifying the obstacles. |
| Keywords: | networks of ART networks of BP fuzzy preceding operation clustering time-shared training |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |