关键词:模糊神经网络;扩展卡尔曼滤波;自组织学习
Fast self-organizing learning algorithm based on EKF for fuzzy neural network
ZHOU Shang-bo,LIU Yu-jiong
(College of Computer Science, Chongqing University, Chongqing 400044, China)
Abstract:To construct an effective fuzzy neural network, this paper presented a self-organizing learning algorithm based on extended Kalman filter for fuzzy neural network. In the algorithm, the network grew rules according to the proposed growing criteria without pruning, speeding up the online learning process.All the free parameters were updated by the extended Kalman filter approach and the robustness of the network was obviously enhanced. The simulation results show that the proposed algorithm can achieve fast learning speed, high approximation precision and generation capability.
Key words:fuzzy neural network; extended Kalman filter(EKF); self-organizing learning
4 结束语
SFNN采用在线学习方法、参数估计和结构辨识同时进行,提高了网络的学习速度。基于该方法生成的模糊神经网络具有紧凑的结构,网络结构不会持续增长,避免了过拟合及过训练现象,确保了系统的泛化能力。
参考文献:
[1]
HUANG Huan,WU Cong-xin.Approximation capabilities of multilayer fuzzy neural networks on the set of fuzzy-valued functions[J].Information Sciences,2009,179(16):2762-2773.
[2]DENG Xing-sheng,WANG Xin-zhou.Incremental learning of dynamic fuzzy neural networks for accurate system modeling[J].Fuzzy Sets and Systems,2009,160(7):972-987.
[3]韦玉科,汪仁煌,李江平,等.一种新的数据智能化处理算法[J].计算机应用研究,2008,25(5):1328-1329.
[4]CHEN Sheng,HONG Xia,LUK B L,et al.Orthogonal-least-squares regression:a unified approach for data modeling[J].Neurocompu-?ting,2009,72(10-12):2670-2681.
?[5]CHEN S,COWAN C F N,GRANT P M.Orthogonal least squares learning algorithm for radial basis function networks[J].IEEE Trans on Neural Networks,1991,2(2):302-309.
[6]CHO K B,WANG B H.Radial basis function based adaptive fuzzy systems and their applications to system identification and prediction[J].Fuzzy Sets and Systems,1996,83(3):325-339.
[7]RIVALS I,PERSONNAZ L.A recursive algorithm based on the extended Kalman filter for the training of feedforward neural models[J].Neurocomputing,1998,20(1):279-294.
[8]SIMON D.Training radial basis neural networks with the extended Kalman filter[J].Neurocomputing,2002,48(1):455-475.
[9]WU Shi-qian,ER M J,GAO Yang.A fast approach for automatic ?generation of fuzzy rules by generalized dynamic fuzzy neural networks[J].IEEE Trans on Fuzzy Systems,2001,9(4):578-594.
[10]WU Shi-qian,ER M J.Dynamic fuzzy neural networks:a novel approach to function approximation[J].IEEE Trans on Systems, Man and Part B-Cybernetics,2000,30(2):358-364.
[11]RIGATOS G,ZHANG Q.Fuzzy model validation using the local statistical approach[J].Fuzzy Sets and Systems,2009,160(7):882-904.