实现了对10个数字音的识别，包含优化类的几个简单示例程序，计算多重分形非趋势波动分析，采用加权网络中节点强度和权重都是幂率分布的模型，针对EMD方法的不足,对信号进行频谱分析及滤波。-To achieve the recognition of 10 digital sound, Optimization class contains several simple sample programs, Calculate the multifractal trend fluctuation analysis, Using weighted model nodes in the network strength and weight are power law distribution, For lack of EMD, The signal spectral analysis and filtering.