| Powell-Beale重开始算法在水质评价中的应用 |
| |
| 引用本文： | 周慧,王晓光,周毅.Powell-Beale重开始算法在水质评价中的应用[J].沈阳理工大学学报,2009,28(4):86-89. |
| |
| 作者姓名： | 周慧 王晓光 周毅 |
| |
| 作者单位： | 1. 沈阳理工大学,理学院,辽宁,沈阳,110168
2. 锦州市环境保护科学研究所,辽宁,锦州,121001 |
| |
| 摘 要： | 建立了基于Powell-Beale重开始算法的改进BP网络水质综合评价模型.该模型在网络权值训练中,以共轭梯度方向取代了标准BP网络的负梯度搜索方向,并通过设定的重新开始准则,适时调整搜索方向,加快了网络运行速度.实验结果表明,新模型改善了标准BP网络模型收敛速度较慢、易陷入局部极小或中断训练的缺陷,泛化能力较强,评价结果客观合理..
|
| 关 键 词： | BP神经网络 Powell-Beale算法 水质评价 随机样本 泛化能力 |
Application of Powell-Beale's Restart Algorithm for Evaluation of Water Quality |
| |
| ZHOU Hui,WANG Xiao-guang,ZHOU Yi.Application of Powell-Beale's Restart Algorithm for Evaluation of Water Quality[J].Transactions of Shenyang Ligong University,2009,28(4):86-89. |
| |
| Authors: | ZHOU Hui WANG Xiao-guang ZHOU Yi |
| |
| Affiliation: | 1.Shenyang Ligong University;Shenyang 110168;China;2.Environmental Protection Science Institute of Jinzhou;Jinzhou 121001;China |
| |
| Abstract: | An Improved BP neural network model based on Powell-Beale's restart algorithm for comprehensive evaluation of water quality is established.In the training of network weights the new model replaces negative gradient search direction in conventional BP network with conjugate gradient search direction and adjusts the search direction timing by the set restart criterion.Thus it can accelerate the network's running speed.The experiment results show that the new model can improve conventional BP network model whi... |
| |
| Keywords: | BP neural network Powell-Beale algorithm water quality evaluation random samples generalization |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
