BP神经网络构建根据拟合非线性函数特点确定BP神经网络结构，由于该非线性函数有两个输入参数，一个输出参数，所以BP神经网络结构为2—5—1，即输入层有2个节点，隐含层有5个节点，输出层有1个节点。　从非线性函数中随机得到2000组输入输出数据，从中随机选择1900组作为训练数据，用于网络训练，100组作为测试数据，用于测试网络的拟合性能。(Construction of BP neural network for nonlinear function fitting structure according to the characteristics of BP neural network, the nonlinear function of two input parameters and one output parameter, so the structure of the BP neural network is 2 - 5 - 1, namely, the input layer has 2 nodes, 5 nodes in hidden layer, the output layer has 1 nodes. 2000 sets of input and output data are obtained randomly from the nonlinear function. 1900 groups are randomly selected as training data and used for network training. The 100 set is used as test data to test the network's fitting performance.)