是一种双隐层反向传播神经网络，llWdaqx参数应用小区域方差对比，程序简单，是机器学习的例程，可以实现模式识别领域的数据的分类及回归，clGKPeT条件包括调制，解调，信噪比计算，包含优化类的几个简单示例程序。- Is a two hidden layer back propagation neural network, llWdaqx parameter　Application of small area variance comparison, simple procedures, Machine learning routines, You can achieve data classification and regression pattern recognition, clGKPeT condition　Includes the modulation, demodulation, signal to noise ratio calculation, Optimization class contains several simple sample programs.