期刊文献+

基于EEMD-Xgboost组合模型的渭北流域月径流序列模拟研究 被引量:5

Simulation of Monthly Runoff Sequences in the Weibei Basin Using Combined EEMD-Xgboost Model
下载PDF
导出
摘要 为提升月径流序列的模拟精度,利用集合经验模式分解(EEMD)析出原始流量数据的模态分量,将极限梯度下降(Xgboost)作为预测函数,构建了基于EEMD-Xgboost的月径流预测模型,并应用EEMD-Xgboost模型训练了黄土坮塬区漆水站1951~1996年月径流序列变化规律,预测了1997~2020年的月径流量。结果表明,与单一Xgboost模型相比,EEMD-Xgboost模型的Nash-Sutcliffe效率(N;)提升了20.27%、均方根误差(R;)减小了93.23%;且EEMD-Xgboost模型优于EEMD-ELM、EEMD-RF模型(N;分别提高2.30%、3.49%;R;减小2.64%、11.75%)。EEMD-Xgboost混合模型集合了数据自适应分析与非线性映射的优点,改善了传统单一模型的预测能力。 To enhance the simulation accuracy of monthly runoff series, an ensemble empirical mode decomposition(EEMD) was used to analyze the modal components of the original flow data. The monthly runoff prediction model was established by taking Xgboost as prediction function. Using the EEMD-Xgboost model, the variation rules of monthly runoff series during 1951-1996 at Chishui station in the Loess Plateau area were trained to predict the monthly runoff from 1997 to 2020. Compared with the single Xgboost model, the results show that the Nash-Sutcliffe efficiency of the EEMD-Xgboost model was improved by 20.27%, and the root mean square error was reduced by 93.23%;The EEMD-Xgboost model outperformed the EEMD-ELM and EEMD-RF models(Nash-Sutcliffe efficiency improved by 2.30%, 3.49%;root mean square error decreased by 2.64%, 11.75%). The hybrid EEMD-Xgboost model gathered the advantages of data adaptive analysis and nonlinear mapping, which improved the prediction ability of the traditional single model.
作者 李蕾 LI Lei(Shaanxi Railway Institute,Weinan 714000,China)
出处 《水电能源科学》 北大核心 2022年第5期22-25,共4页 Water Resources and Power
基金 渭北流域径流序列模拟研究课题(KY2021-51)。
关键词 EEMD模态分解 模态分量 Xgboost模型 机器学习 径流量预测 EEMD modal decomposition modal component Xgboost model machine learning runoff prediction
  • 相关文献

同被引文献52

引证文献5

二级引证文献6

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部