【摘要】 本文通过对多层神经网络模型的BP算法的推导,讨论用BP算法对网络进行优化后误差精度、网络结构和学习次数之间的关系,最后指出BP算法的不足之处,提出加快学习过程的一些改进方法和技巧。更多还原

【Abstract】 This paper drew an inference of the BP back propagation from the multi-layer neural networks in details. Then the author discussed the relationships among the relative accuracy network structure and the learning frequency after the neural networks were in optimization. In the end the author pointed out that the defects of the BP in the multi- layer neural networks and brought up some key improving methods and techniques to speed up the learning process.更多还原