摘要
ＢＰ学习算法是处理非线性优化问题的一种新方法，其网络配置的不同直接影响网络收敛的速度与精度．本文系统地分析了一些对建立时序预测模型产生影响的因素。
BP learning algorithm is a new method for processing nonlinear optimization problems. But Back Propagation is too slow convergence rates for inappropriate configuration. This paper systematically analyzed some factors which affected neural network convergence rates, and a optimization network training method is designed.
出处
《郑州大学学报（自然科学版）》
1998年第2期23-26,39,共5页
Journal of Zhengzhou University (Natural Science)
关键词
神经网络
BP算法
优化配置
时序建模
预测
neural network
BP algorithm
Configuration
prediction and modeling
