摘要
将模糊C-均值(FCM)算法和模糊对向传播神经网络相结合,解决了模糊对向传播神经网络中模糊隶属度函数的自动生成问题,在此基础上结合TSK(SugenoTanakaKang)模糊模型,提出了改进的模糊对向传播神经网络ImprovedFCP,并给出了学习算法。对两种典型的非线性模型进行实验研究,实验结果表明ImprovedFCP网络具有良好的非线性逼近能力。
Fuzzy Cmean (FCM) algorithm and fuzzy counter propagation neural network (FCP) are combined in this paper. The problem, automatic generation of fuzzy membership function in FCP, has been solved. An improved fuzzy counter propagation neural network (improved FCP/IFCP) and the learning algorithm of IFCP have been proposed based on the TSK model. The experiments have been made on the two typical nonlinear models. The experimental results show that the improved FCP enjoys the good performance of nonlinear approximation capability.
出处
《华东船舶工业学院学报》
2003年第2期45-49,共5页
Journal of East China Shipbuilding Institute(Natural Science Edition)
基金
江苏省自然科学基金资助项目(BK2002001)
