摘要
BP网络是广泛应用的神经网络之一,但其收敛速度慢、易陷入局部最小。针对BP神经网络这些问题,本文提出了归一化输出误差的方法,引入了网络输出对权值的影响,有效地提高了网络的训练时间。仿真结果表明此种方法对改善网络的收敛速度效果明显,且对于其它的BP网络改进算法也有较好的改善作用。
The back-propagation(BP) neural network is one of the most widely applied neural networks,but it has some disadvantages, including slowly learning convergent velocity and easily converging to local minimum.In this paper,the method of normalized output error is proposed.The results of computer simulation prove that the new method is effective to solve some problems and faster than the traditional methods in training multi-layer feed-forward neural networks,and it also can improve the training time of other methods .
出处
《微计算机信息》
北大核心
2006年第07S期305-306,53,共3页
Control & Automation
基金
北京市自然科学基金重点项目项目编号:4031001-1
关键词
神经网络
BP算法
训练时间
Neural network, BP algorithm, Training time