摘要
本文研究了基本BP算法的工作原理,分析了导致基本BP算法学习效率低的原因,提出了基于附加动量法的改进BP算法。经过理论分析和实验验证,在基本BP算法中加入附加动量的方法,可以加快BP神经网络的学习速度,提高BP神经网络的学习效率。
The basic BP algorithm work principle is discussed and the reason for low efficiency of basic BP algorithm is analyzed. Improved BP algorithm based on additional momentum is proposed. Through theoretical analysis and experimental verification, adding additional momentum in the basic BP algorithm can accelerate the learning speed of BP neural network and advance the learning efficiency.
出处
《济源职业技术学院学报》
2012年第3期9-13,共5页
Journal of Jiyuan Vocational and Technical College
基金
河南省教育厅青年骨干教师资助项目(2010GGJS-259)
关键词
BP算法
最速下降法
学习效率
附加动量法
BP algorithm
gradient method
learning efficiency
additional momentum
