该程序是用C语言写的采用了动量方法的BP神经网络，它具有极其强大的扩展性，可用于函数的逼近及分类识别等各方面，语言书写规范，其中大量使用了动态分配一、二维数组的方法，该网络实例中实现的是对某一函数的逼近，输入层为二层，中间层可自己设定多达十几层，逼近精度非常精确。读者可根据自己的需要设定各层的数目以及相应的训练样本进行训练与测试。-The program is written in C language using the momentum method of BP neural network, which has extremely powerful scalability, can be used for function approximation and classification in areas such as language specification, including large-scale use of a dynamically allocated one, two dimensional array of methods, examples of the network realize that the approximation of a function, for the second input layer, middle layer can be set as many as a dozen layers, close to very precise accuracy. Readers may need to be set in accordance with its own number of floors and the corresponding training samples for training and testing.