摘要
文中提出了将模糊聚类与梯度算法相结合的一种改进的训练模糊神经网络的混合型算法．模拟结果表明，模糊神经网络可以成功地用于时间序列的预测，模糊神经网络的训练速度与模拟精度都优于传统多层ＢＰ网络．
An improved learning algorithm is proposed,which integrates fuzzy cluster with gradient algorithm to train fuzzy neural networks.The simulated results show that fuzzy neural networks can be successfully applied to the prediction of time series and that both training speed and simulated precision of fuzzy neural networks are better than those of traditional multilayer BP networks.
出处
《计算机研究与发展》
EI
CSCD
北大核心
1998年第7期663-667,共5页
Journal of Computer Research and Development
基金
国家自然科学基金
国家教委符号计算与知识工程开放研究实验室资助
关键词
模糊系统
时间序列预测
神经网络
artificial neural networks, fuzzy systems, time series analysis,fuzzy cluster, mixed training algorithm
