摘要
提出了一种新的训练多层神经网络的适应性共轭梯度学习算法。理论分析和仿真结果证明 ,它在模式识别领域里是—种可行和有效的算法 ,而且其识别能力优于传统的BP算法 。
REHEMUJIANG, GULI\5TURSUN, ANWAR\5TURDI, et al. Artificial Intelligence Center in the University of Petroleum, China, Beijing 102249 A new kind of adaptive conjugate gradient algorithm for training multi-layer neural networks was proposed. The theoretical analysis and simulation results proved that it is an effective algorithm in the field of pattern recognition. It has more powerful and higher recognition ability and convergence speed than the conventional BP algorithm.
出处
《石油大学学报（自然科学版）》
CSCD
北大核心
2003年第6期108-111,共4页
Journal of the University of Petroleum,China(Edition of Natural Science)
关键词
多层神经网络
BP算法
共轭梯度法
图像识别
neural network
BP algorithm
conjugate gradient method
image recognition
