此程序是神经网络中基于梯度的径向基函数算法，在MATLAB中实现。用一个2－n－1结构的RBF网对SISO系统进行建模，网络的两个输入为u(k-1)和y(k-1)，输出为 y(k)。令y(0)=0，按飞线性系统产生200个样本，其中前100个样本用于训练，后100个样本用于测试。-This procedure is based on the gradient neural network radial basis function algorithm is implemented in MATLAB. With a 2-n-1 structure of RBF network model of the SISO system, the network of two input u (k-1) and y (k-1), output y (k). So that y (0) = 0, the linear system produced by flying 200 samples, of which the first 100 samples for training, after 100 samples for testing.