首页 | 本学科首页   官方微博 | 高级检索  
     

模糊神经网络的快速学习算法研究
引用本文:刘俊强,伞冶,王子才.模糊神经网络的快速学习算法研究[J].高技术通讯,2000,10(6):48-50,21.
作者姓名:刘俊强  伞冶  王子才
作者单位:哈尔滨工业大学仿真中心,哈尔滨,150001
基金项目:国家自然科学基金资助项目!( 698740 0 7)
摘    要:对模糊神经网络的快速学习算法进行了研究,提出FNN的最小二乘-单纯形产算法。解决了FNN采用梯度型算法学习时所固有的局部收敛问题及跗算法在线应用的问题。与梯度型算法以及GA相比,LS-Simplex算法具有更高的学习精度和更快的收敛速度,所得的FNN具有良好的泛化性能。

关 键 词:模糊神经网络  学习算法  最小二乘法

A Least Square-Simplex Algorithm for Training Fuzzy Neural Networks
Liu Junqiang,San Ye,Wang Zicai.A Least Square-Simplex Algorithm for Training Fuzzy Neural Networks[J].High Technology Letters,2000,10(6):48-50,21.
Authors:Liu Junqiang  San Ye  Wang Zicai
Abstract:A novel fast algorithm named Least Square-Simplex (LS-Simplex), for training f uz zy neural networks(FNN) is proposed.The algorithm obtains the performance of glo bal convergence and avoids the inherent local convergence when adopting grads al gorithm to train FNN. Also it accelerates FNN's training and can be used on-lin e which is impossible when using genetic algorithm(GA). Compared with grads algor ithm and GA, the LS-Simplex owns more accurate precision and faster convergent speed, and the FNN obtained has excellent performance of generalization.
Keywords:Fuzzy neural networks  Training algorithm  Least square technique  Si mplex search algorithm
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号