基于LSFD算法的小波模糊神经网络的研究与应用  

Research and Application of Wavelet Fuzzy Neural Network Based on LSFD Algorithm

作　　者：邵俊倩[1] 

机构地区：[1]绥化学院信息工程学院,绥化152061

出　　处：《计算机与数字工程》2013年第9期1423-1425,1432,共4页Computer & Digital Engineering

基　　金：黑龙江教育厅科学技术研究项目(编号:12533077);绥化学院2012科学技术研究资助项目

摘　　要：利用小波变换的多分辨率特性构造小波模糊神经网络模型,并应用在非线性系统的辨识上。在参数学习上,给出了模糊微分与李亚普诺夫稳定相结合的新算法—LSFD算法,并与梯度下降法进行了对比。通过仿真,结果表明小波模糊神经网络模型与模糊神经网络、模糊小波神经网络、小波神经网络和神经网络等模型相比,其性能指标最小,收敛速度更快,更加准确。Using multi-resolution characteristics of the wavelet, the model of wavelet fuzzy neural network is constructed, it is applied for the identification of non-linear system. The new learning algorithm of a combination of Lyapunov stability and fuzzy differential is given, a comparison is done with gradient descent. Compared with models of fuzzy neural network, fuzzy wavelet neural network, wavelet neural network and neural network, the simulations to non-linear system demonstrate that the wavelet fuzzy neural network has the performance index of the minimum, the faster convergent speed, and more accurate.

关 键 词：T-S模型 小波模糊神经网络 模糊微分 李亚普诺夫稳定 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]