Short-term streamflow time series prediction model by machine learning tool based on a data preprocessing technique and swarm intelligence algorithm,Hydrological Sciences Journal

当前位置： X-MOL 学术 › Hydrol. Sci. J. › 论文详情

Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)

Short-term streamflow time series prediction model by machine learning tool based on a data preprocessing technique and swarm intelligence algorithm
Hydrological Sciences Journal ( IF 3.5 ) Pub Date : 2020-10-30 , DOI: 10.1080/02626667.2020.1828889
Wen-jing Niu ₁ , Zhong-kai Feng ₂ , Wen-fa Yang ₁ , Jun Zhang ₁

Affiliation

ABSTRACT Accurate streamflow prediction information is of great importance for water resource planning and management. The goal of this research is to develop a hybrid model for forecasting short-term runoff time series, where the variational mode decomposition (VMD) is first used to decompose the original nonlinear natural streamflow into numerous subcomponents with different frequencies and resolutions. Second, the extreme learning machine (ELM) is used to excavate the complicated input–output relationship hidden in each subcomponent, and the emerging sine cosine algorithm (SCA) is used to determine the suitable network parameter for each ELM model. Finally, the forecasting results of all the modelled subcomponents are summarized to form the forecasting result for original streamflow. Based on several statistical evaluation measures, the feasibility of the hybrid method is investigated in runoff forecasting for Danjiangkou Reservoir in China. The results indicate that the hybrid method can produce superior forecasting results compared to several control methods, providing accurate streamflow prediction information for operators.

更新日期：2020-10-30

点击分享查看原文

点击收藏

阅读更多本刊最新论文

全部期刊列表>>