摘要
提出具有动态最优学习率的ＢＰ算法， 动态最优学习率是通过迭代的方式得到的， 这种算法具有运算简单，收敛速度快，精度高特点。仿真结果说明此算法是有效的。
This paper considers an efficient back propagation algorithm using dynamic optimal learning rate. The optimal learning rate is obtained by an iteration formula. This algorithm has the following merits: simplicity of computation, quickness of convergence and high precision. The simulation results are given to demonstrate the effectiveness of this algorithm.
出处
《系统仿真学报》
EI
CAS
CSCD
2001年第6期775-776,815,共3页
Journal of System Simulation
关键词
BP算法
多层前馈神经网络
动态最优学习率
back propagation algorithm
Multiplayer
feed forward
neural network
dynamic optimal learn rate
