| 基于模糊神经网络结构的多值联想记忆 |
| |
| 引用本文： | 舒桂清,肖平. 基于模糊神经网络结构的多值联想记忆[J]. 计算机工程与应用, 2002, 38(2): 94-96 |
| |
| 作者姓名： | 舒桂清 肖平 |
| |
| 作者单位： | 广东省科技干部学院计算机与电子工程系,广州,510640;华南理工大学电子与通讯工程系,广州,510641 |
| |
| 摘 要： | 文章基于模糊神经网络结构,即通过模糊化,推理,去模糊三个过程,把Kosko提出的模糊联想记忆(FAM)网络模型应用到容错性需要较强的多值联想记忆中,解决了这种网络模型不能对随机噪声模式正确联想的问题,新的网络模型设计简单,大量实验表明文中的联想记忆网络大大提高了FAM网络的容错性能。
|
| 关 键 词： | 模糊神经网络 联想记忆 模糊推理 |
| 文章编号： | 1002-8331-(2002)02-0094-03 |
| 修稿时间： | 2000-12-01 |
Multi-value Associative Memory Based on the Structure of Fuzzy Neural Networks |
| |
| Shu Guiqing Xiao Ping. Multi-value Associative Memory Based on the Structure of Fuzzy Neural Networks[J]. Computer Engineering and Applications, 2002, 38(2): 94-96 |
| |
| Authors: | Shu Guiqing Xiao Ping |
| |
| Affiliation: | Shu Guiqing 1 Xiao Ping 21 |
| |
| Abstract: | In this paper,on the basis of scheme of fuzzy neural network,i.e.,with the processes of fuzziness,inference,defuzziness,fuzzy associative memories(FAM)proposed by Kosko are applied to the multi-value associative memories,which are needed to be higher error-tolerance to the random noise patterns.So the new scheme solves the problem that the FAM network cannot be recalled correctly to the random noise patterns,The new model is easy to be designed.A large number of experiments show the new FAM proposed by this paper have a stronger capability of error-tolerance than that of the Kosko's. |
| |
| Keywords: | fuzzy neural networks associative memory fuzzy inference |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
