-----引入动量的算法

建立一个3层（含输入层）的BP神经网络，并对其进行训练

输入层不进行数据处理，隐含层激活函数为sigmod函数，输出层为线性函数

输入输出数据归一化到[-1,1]，数据在矩阵中按行向量表示

即x=[x11,x12 x21,x22 ... xp1,xp2] y=[y1 y2 ... yp] p为样本数 -The introduction of the algorithm ----- momentum to build a three-layer (including input layer) of the BP neural network, and its training　of the input layer does not carry out data processing, the hidden layer activation function sigmod function, the output layer-line Sexual Function　input and output data normalized to [-1,1], the data in the matrix row-vector representation　or x = [x11, x12　x21, x22　...　xp1, xp2] y = [y1　 y2　...　yp] p as the number of samples