[1]胡彩虹,王艳菊,吴泽宁..基于聚类的支持向量机在洪水预报中的应用[J].郑州大学学报(工学版),2009,30(04):123-127.[doi:10.3969/j.issn.1671-6833.2009.04.030]
　Hu Caihong,Wang Yanju,Wu Zening.Application of clustering-based support vector machine in flood forecasting[J].Journal of Zhengzhou University (Engineering Science),2009,30(04):123-127.[doi:10.3969/j.issn.1671-6833.2009.04.030]

点击复制

基于聚类的支持向量机在洪水预报中的应用()

分享到：

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:

30卷

期数:

2009年04期

页码:

123-127

栏目:

出版日期:

1900-01-01

文章信息/Info

Title:

Application of clustering-based support vector machine in flood forecasting

作者:

胡彩虹; 王艳菊; 吴泽宁.

郑州大学水利与环境学院,河南郑州,50001, 郑州大学水利与环境学院,河南郑州,50001, 郑州大学水利与环境学院,河南郑州,50001

Author(s):

Hu Caihong; Wang Yanju; Wu Zening

关键词:

系统聚类;  支持向量机;  洪水预报

Keywords:

systematic clustering;  support vector machines;  Flood forecasting

DOI:

10.3969/j.issn.1671-6833.2009.04.030

文献标志码:

A

摘要:

半干旱地区的特殊特点使其径流模拟计算难度增大,且难以获得较详细的资料,因而洪水预报难度大,尤其是洪峰流量的预报.若应用所有样本进行模型参数确定并预报,不能完全反映洪水的不同特性.因此采用了基于聚类分析的支持向量机模型,以半干旱半湿润地区的岚河流域为例,进行了模拟检验,结果表明,效率系数大部分达到85%以上,平均相对误差绝对值多数都小于1.5%.另外洪峰流量相对误差绝对值均在15%以内,特别洪峰流量较大的几场洪水,相对误差小于1%.洪峰流量和峰现时差合格率均达100%.

Abstract:

The special characteristics of semi-arid areas make runoff simulation and calculation more difficult, and it is difficult to obtain more detailed data, so flood forecasting is difficult, especially the prediction of peak flood flow. If all samples are used to determine and predict model parameters, it cannot fully reflect the different characteristics of floods. Therefore, the support vector machine model based on cluster analysis is adopted, and the simulation test is carried out in the Lanhe River Basin in the semi-arid and semi-humid area, and the results show that most of the efficiency coefficients reach more than 85%, and most of the absolute values of the average relative error are less than 1.5%.In addition, the absolute relative error values of the peak flow are within 15%, especially the relative error of several floods with large peak flow is less than 1%.The pass rate of peak flow and peak current time difference reaches 100%.

常用功能