Design of RBF Network Based on MATLAB
Systemic dynamic modeling and control are always the important fields of scientific research. For technical complicacy, to get the accurate mathematical modeling is somewhat difficult .Neural networks is the one aspect of the intellective technology which develops promptly. Thereinto RBF (Radial Basic Function) neural networks has been widely applied to nonlinear modeling, function approach, systemic recognition and so on . As to actual problem, there were no system method to solve network architecture and neurons, quite a few experiments must be made. The paper mostly introduces the way to create a neural network ,train a neural network and simulate a neural network with MATLAB language on the basic of synthetically referring the theory of neural networks .The way makes two grade transmittable arithmetic be the main body of RBF neural networks arithmetic , in another words , to make the first grade transmittable arithmetic confirm the concealed layer figures of RBF neural network and then to make the second grade transmittable arithmetic confirm the basic function parameter of RBF neural network . The paper gives the examples and simply explanation of the proceeding which has to be noticed during the design. This network has many merits, such as quick rate of learning, better capability of approach and so on.
RBF network Neural network MATLAB language
Chang xiaoli
North University of China ,Shan Xi ,TaiYuan ,030051,China
国际会议
北京
英文
2007-08-05（万方平台首次上网日期，不代表论文的发表时间）