| 基于Bagging算法的水位预测模型 |
| |
| 引用本文： | 耿悦敏.基于Bagging算法的水位预测模型[J].广东技术师范学院学报,2007(10):28-30. |
| |
| 作者姓名： | 耿悦敏 |
| |
| 作者单位： | 广东交通职业技术学院,广东广州510650 |
| |
| 摘 要： | 传统的统计分析在小样本预测中的效果不佳,虽然神经网络一定程度上解决了传统方法所遇到的问题,但样本的数量又影响了神经网络的泛化能力,神经网络集成的方法较好地解决了这一问题.在运用智能计算技术建立BP网络的基础上,再利用Bagging算法构造神经网络的集成模型.用珠江三角洲天河水文站的数据进行训练和预测,结果表明,基于Bagging算法的神经网络集成的预测不仅解决了样本数据量少、偏差大、不确定性的问题,而且比单一神经网络具有更强的泛化能力,预测更为可靠.
|
| 关 键 词： | Bagging算法 神经网络集成 水位预测 |
| 文章编号： | 1672-402X（2007）10-0028-03 |
| 收稿时间： | 2007-09-10 |
| 修稿时间： | 2007年9月10日 |
Forecast Water Lever Model based on Bagging Algorithm |
| |
| Geng Yue-min.Forecast Water Lever Model based on Bagging Algorithm[J].Journal of Guangdong Polytechnic Normal University,2007(10):28-30. |
| |
| Authors: | Geng Yue-min |
| |
| Institution: | Guangdong Communications Polytechnic, Guangzhou 510650, China |
| |
| Abstract: | The traditional statistical analysis is not good in the small sample forecast effect; although nerve network in the certain degree has solved the problem which the traditional method meets, the sample quantity affected the nerve network to exude the ability; the network ensemble method has solved this problem well. Using the intelligence computa- tion the BP network has been established, then structuring network ensemble model with the Bagging algorithm. Carrying on the training and the forecast with the Tianhe hydrometric Station of Pearl River Delta. data, the result indicated that, the forecast with Bagging-based neural network ensemble model has better generalization ability and more reliability than sole neural network, especially when dealing with the problem of small sample, biggish deviation and uncertainty. |
| |
| Keywords: | bagging algorithm neural network ensemble water lever forecasting |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
