构成不同频率的调制信号，采用加权网络中节点强度和权重都是幂率分布的模型，使用拉亚普诺夫指数的公式，BP神经网络用于函数拟合与模式识别，基于负熵最大的独立分量分析。- Constituting the modulated signals of different frequencies, Using weighted model nodes in the network strength and weight are power law distribution, Raya Punuo Fu index using the formula, BP neural network function fitting and pattern recognition, Based on negative entropy largest independent component analysis.