1、从训练集中取出某一样本，把该样本的信息数据输入网络中。2、通过各节

1，从训练集中提取的样本中，放入该样本的信息数据输入网络中<br>。2，通过各个互连之间的连接情况正向逐层处理后，得到神经网络的实际输出<br>。3，计算网络实际输出与期望输出的误差值。<br>如图4所示，将误差逐层反向回传至之前各层，并按一定原则将误差信号加载到连接权值上，使整个神经网络的连接权值向误差减小的方向转化。<br>5，対训练集中每一个输入-输出样本重复以上步骤，直到整个训练样本集的误差减小到符合要求为止。

1、从训练集中取出某一样本,把该样本的信息数据输入网络中。<br>2、通过各节点间的连接情况正向逐层处理后,得到神经网络的实际输出。<br>3、计算网络实际输出与期望输出的误差值。<br>4、将误差逐层反向回传至之前各层,并按一定原则将误差信号加载到连接权值上,使整个神经网络的连接权值向误差减小的方向转化。<br>5、対训练集中每一个输入—输出样本重复以上步骤,直到整个训练样本集的误差减小到符合要求为止。

1. Take a sample from the training set and input the information data of the sample into the network.<br>2. The actual output of the neural network is obtained after the connection between the nodes is processed layer by layer.<br>3. Calculate the error value between the actual output and the expected output of the network.<br>4. The error is transmitted back to the previous layers layer by layer, and the error signal is loaded on the connection weight according to certain principles, so that the connection weight of the whole neural network is transformed to the direction of error reduction.<br>5、 Repeat the above steps for each input-output sample in the training set until the error of the whole training sample set is reduced to meet the requirements.<br>