数据误差处理方法在洪水预报中的应用
Application of the Data Error Processing Method in Flood Prediction
丁海蛟 1车文刚1
作者信息
- 1. 昆明理工大学信息工程与自动化学院,云南昆明650500
- 折叠
摘要
洪水灾害会对我国的正常农业生产造成非常严重的影响,因此提出一种基于数据误差处理的洪水预报方法有重要意义.结合四川省自贡市某水文站实测的流量数据,应用LS-SVM智能算法建立了单输入单输出(流量—流量)洪水预报模型,并应用数据误差处理方法中改进的拉依达准则法(3σr)和肖维勒准则法(Chauvenet)来处理样本数据里存在的一些误差数据.实验结果表明,应用以上2种数据误差处理方法处理过后的LS-SVM预测模型可以满足洪水预报精度的要求.
Abstract
Since the flood disaster causes very serious impacts on normal agricultural production in China,it was of great significance to put forward the flood prediction method based on data error processing.Combining with the flow data measured by a given hydrological station in Zigong City of Sichuan Province,LS-SVM intelligence algorithm was used to establish a single input single output (flow-flow) prediction model.Two kinds of the data error processing method:the improved 3σ method and the chauvenet method to deal with some error data which were hiding in the sample data.The experimental results showed that the application of the above two kinds of data error treatment method used in the LS-SVM prediction model could meet the requirements of flood prediction accuracy.
关键词
流量/LS-SVM/洪水预报/数据误差处理Key words
Flow/LS-SVM/Flood prediction/Data error processing引用本文复制引用
出版年
2016
国家科技期刊平台
NSTL
维普
万方数据