论文部分内容阅读
本文介绍了传统的BP算法,分析了它的不足之处,研究了一种加快网络收敛速度的基于共轭梯度法的改进BP算法。通过实例,运用MATLAB语言进行了仿真,结果表明改进算法的学习收敛速度大大优于标准BP算法。
This paper introduces the traditional BP algorithm and analyzes its shortcomings. An improved BP algorithm based on conjugate gradient method is proposed to speed up the network convergence. Through the example, using MATLAB language simulation, the results show that the improved algorithm learning convergence rate is much better than the standard BP algorithm.