首页 | 本学科首页   官方微博 | 高级检索  
     

一种基于双曲函数的BP网络改进算法
引用本文:王玲芝.一种基于双曲函数的BP网络改进算法[J].西安邮电学院学报,2009,14(5):93-95,99.
作者姓名:王玲芝
作者单位:西安邮电学院计算机科学与技术系,陕西,西安,710121
摘    要:提出了一种新的BP(Back Propagation)网络权值调整规则算法。该算法以双曲函数为基础,将BP网络输出层节点的实际输出值与期望输出值的差作为双曲函数的输入,输出为权值的变化量。实例仿真表明,该算法简单,容易实现,加快了网络的学习速度,提高了学习效率。

关 键 词:BP网络  权值调整  双曲函数

An improved algorithm based on hyperbolic function of BP neural network
WANG Ling-zhi.An improved algorithm based on hyperbolic function of BP neural network[J].Journal of Xi'an Institute of Posts and Telecommunications,2009,14(5):93-95,99.
Authors:WANG Ling-zhi
Affiliation:WANG Ling - zhi (Department Of Computer Science, Xi'an University of Posts and Telecommunications, Xi'an 710121, China)
Abstract:An improved algorithm of BP(Back Propagation) neural network is proposed, and a new weight- adjusting rule is introduced in the paper. The algorithm is based on hyperbolic function whose input is the error between actual output value of the output layer node and the expected output value, and the output is a variation value of weight. The simulation of instances show that the algorithm is simple and easy to implement. Meanwhile, it quickens the study of the network, and improves the efficiency.
Keywords:BP neural network  weight- adjusting  hyperbolic function
本文献已被 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号