| 基于鲁棒PCA聚类分析的径流周期特性研究 |
| |
| 引用本文： | 邓红霞,李存军,朱兵,丁晶.基于鲁棒PCA聚类分析的径流周期特性研究[J].武汉大学学报(工学版),2006,39(1):10-14. |
| |
| 作者姓名： | 邓红霞 李存军 朱兵 丁晶 |
| |
| 作者单位： | 1. 四川大学水电工程学院,四川,成都,610065
2. 四川大学建筑与环境学院,四川,成都,610065 |
| |
| 摘 要： | 为了研究多年来河流天然径流的变化周期,引入鲁棒PCA(Principal Component Analysis)方法实现大量非线性流量数据的压缩和特征提取,得到了更具表征性的主成分,在此基础上提出了基于加权主成分距离和密度的径流聚类分析方法和径流周期特性分析的模糊模式识别算法.通过嘉陵江支流白龙江下游的三磊坝水文站1964~2002年周平均流量数据的实验和仿真,得出了该站径流具有7年的显著周期和相应的隶属度,从而证明了本文提出的基于鲁棒PCA聚类分析的径流周期特性研究方法具有较好的适用性.
|
| 关 键 词： | 径流周期特性分析 鲁棒PCA 聚类分析 模式识别 |
| 文章编号： | 1671-8844（2006）01-010-05 |
| 收稿时间： | 2005-06-28 |
| 修稿时间： | 2005年6月28日 |
Runoff periodicity analysis based on robust PCA and cluster |
| |
| DENG Hongxia,LI Cunjun,ZHU Bing,DING Jing.Runoff periodicity analysis based on robust PCA and cluster[J].Engineering Journal of Wuhan University,2006,39(1):10-14. |
| |
| Authors: | DENG Hongxia LI Cunjun ZHU Bing DING Jing |
| |
| Affiliation: | 1. School of Hydroelectric Eng. , Sichuan Univ. , Chengdu 610065, China;2. School of Architecture and Environment, Sichuan Univ. , Chengdu 610065, China |
| |
| Abstract: | An effective analysis method of fuzzy pattern recognition algorithm based on cluster analysis by distance and density,with the more representative weighted principal component what is reduced and distilled from large numbers of nonlinear original data by robust PCA,is proposed to study periodicity of nature runoff in many years;accordingly weekly average runoff of Sanleiba Hydrology Station in lower Bailong River,a branch of Jialingjiang River,in the period of 1964-2002 year,is tested by this method;resulted in prominent 7-year period with remarkable subjective. Experimental results have shown that the method of runoff periodicity analysis based on robust PCA and cluster runs approving adaptability. |
| |
| Keywords: | runoff periodicity analysis robust PCA cluster analysis pattern recognition |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
