摘要
解决了BP神经网络结构参数、学习速率与初始权值的选取问题,并对传统的BP算法进行了改进,提出了BP神经网络自适应学习算法,又将其编制成计算机程序,使得输入节点、隐层节点和学习速率的选取全部动态实现,减少了人为因素的干预,改善了学习速率和网络的适应能力.计算结果表明:BP神经网络自适应学习算法较传统的方法优越,训练后的神经网络模型不仅能准确地拟合训练值,而且能较精确地预测未来趋势.
We resolve the problem of selecting architectural parameters, learning rate, initial connection weights and improves BP algorithm of artificial neural network. The self\|adapting algorithm of BP artificial neural network has been proposed, and programmed a C language procedure. It can make the selection of input units, hidden units and learning rate easily in the course of training, reduce external interference and improve the adaptive ability of learning rate and neural network. Our conclusion shows that the self\|adapting algorithm of BP artificial neural network superior to the statistical modeling approach and the traditional BP artificial neural network, it can not only exactly imitate training valuation but also make prediction accurately.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2004年第5期1-8,共8页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(70271073)
国家社会科学基金(01BJY025)
关键词
人工神经网络
BP算法
自适应
自组织方法
artificial neural network
BP algorithm
self-adapting
group method of data handling