k型局部连接神经网络的结构和算法，及其与其他网络的函数逼近性能对比-k partially connected neural network structure and algorithm, and other network function approximation performance comparison

解决了非线性函数的逼近问题,实现了BP神经网络在非线性函数逼近中的应用。得出了信号的频率与隐含点之间，隐含点与函数逼近能力之间的关系——隐含层神经元数目越多。-Solved the problem of nonlinear function approximation, and realized the application of the BP neural network in nonlinear function approximation. Between signal frequenc