Runoff Forecasting Research Coupling Quadratic Factor Screening and Deep Learning
The effective screening of factors influencing runoff is a key aspect of runoff forecasting research.However,there are many factors affecting runoff,and these factors have complex interactions.Most of the existing studies use numerically driven models with primary factor screening,and the results sh...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | zho |
| Published: |
Editorial Office of Pearl River
2023-01-01
|
| Series: | Renmin Zhujiang |
| Subjects: | |
| Online Access: | http://www.renminzhujiang.cn/thesisDetails#10.3969/j.issn.1001-9235.2023.06.006 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841536147612762112 |
|---|---|
| author | CHENG Liwen HUANG Shengzhi LI Pei LI Ziyan JIA Songtao HUANG Qiang |
| author_facet | CHENG Liwen HUANG Shengzhi LI Pei LI Ziyan JIA Songtao HUANG Qiang |
| author_sort | CHENG Liwen |
| collection | DOAJ |
| description | The effective screening of factors influencing runoff is a key aspect of runoff forecasting research.However,there are many factors affecting runoff,and these factors have complex interactions.Most of the existing studies use numerically driven models with primary factor screening,and the results show that the input factors are spatially redundant,leading to poor forecasting results.In view of this,the support vector regression (SVR) and the long-short memory network model (LSTM) are compared with Weihe River Basin as an example,and the LSTM model is selected as the optimal forecasting model.Principal component analysis and gray correlation analysis are used for secondary screening of the input terms to form a model coupling principal component analysis,gray correlation analysis,and LSTM.The results show that:①the fitting accuracy of LSTM is higher than that of SVR;②the secondary screening of the input terms improves the forecast accuracy,and the forecast accuracy of the coupled model is better than that of the single model,specifically,the model accuracy evaluation indexes of the coupled model are substantially improved compared with those of the single model;③the Nash efficiency coefficient and deterministic coefficient of the coupled model of gray system correlation analysis are improved by 0.13% and 0.03%,respectively,compared with those of the coupled model of principal component analysis,and the standard deviation ratio of observed values is improved by 42.9%.The study shows that the secondary factor screening by using gray correlation can effectively improve forecast accuracy. |
| format | Article |
| id | doaj-art-fe361560b2504c03b3ca7b8cf4c394f1 |
| institution | Kabale University |
| issn | 1001-9235 |
| language | zho |
| publishDate | 2023-01-01 |
| publisher | Editorial Office of Pearl River |
| record_format | Article |
| series | Renmin Zhujiang |
| spelling | doaj-art-fe361560b2504c03b3ca7b8cf4c394f12025-01-15T02:22:10ZzhoEditorial Office of Pearl RiverRenmin Zhujiang1001-92352023-01-014447637832Runoff Forecasting Research Coupling Quadratic Factor Screening and Deep LearningCHENG LiwenHUANG ShengzhiLI PeiLI ZiyanJIA SongtaoHUANG QiangThe effective screening of factors influencing runoff is a key aspect of runoff forecasting research.However,there are many factors affecting runoff,and these factors have complex interactions.Most of the existing studies use numerically driven models with primary factor screening,and the results show that the input factors are spatially redundant,leading to poor forecasting results.In view of this,the support vector regression (SVR) and the long-short memory network model (LSTM) are compared with Weihe River Basin as an example,and the LSTM model is selected as the optimal forecasting model.Principal component analysis and gray correlation analysis are used for secondary screening of the input terms to form a model coupling principal component analysis,gray correlation analysis,and LSTM.The results show that:①the fitting accuracy of LSTM is higher than that of SVR;②the secondary screening of the input terms improves the forecast accuracy,and the forecast accuracy of the coupled model is better than that of the single model,specifically,the model accuracy evaluation indexes of the coupled model are substantially improved compared with those of the single model;③the Nash efficiency coefficient and deterministic coefficient of the coupled model of gray system correlation analysis are improved by 0.13% and 0.03%,respectively,compared with those of the coupled model of principal component analysis,and the standard deviation ratio of observed values is improved by 42.9%.The study shows that the secondary factor screening by using gray correlation can effectively improve forecast accuracy.http://www.renminzhujiang.cn/thesisDetails#10.3969/j.issn.1001-9235.2023.06.006runoff forecastingfactor screeningSVRLSTMWeihe River Basin |
| spellingShingle | CHENG Liwen HUANG Shengzhi LI Pei LI Ziyan JIA Songtao HUANG Qiang Runoff Forecasting Research Coupling Quadratic Factor Screening and Deep Learning Renmin Zhujiang runoff forecasting factor screening SVR LSTM Weihe River Basin |
| title | Runoff Forecasting Research Coupling Quadratic Factor Screening and Deep Learning |
| title_full | Runoff Forecasting Research Coupling Quadratic Factor Screening and Deep Learning |
| title_fullStr | Runoff Forecasting Research Coupling Quadratic Factor Screening and Deep Learning |
| title_full_unstemmed | Runoff Forecasting Research Coupling Quadratic Factor Screening and Deep Learning |
| title_short | Runoff Forecasting Research Coupling Quadratic Factor Screening and Deep Learning |
| title_sort | runoff forecasting research coupling quadratic factor screening and deep learning |
| topic | runoff forecasting factor screening SVR LSTM Weihe River Basin |
| url | http://www.renminzhujiang.cn/thesisDetails#10.3969/j.issn.1001-9235.2023.06.006 |
| work_keys_str_mv | AT chengliwen runoffforecastingresearchcouplingquadraticfactorscreeninganddeeplearning AT huangshengzhi runoffforecastingresearchcouplingquadraticfactorscreeninganddeeplearning AT lipei runoffforecastingresearchcouplingquadraticfactorscreeninganddeeplearning AT liziyan runoffforecastingresearchcouplingquadraticfactorscreeninganddeeplearning AT jiasongtao runoffforecastingresearchcouplingquadraticfactorscreeninganddeeplearning AT huangqiang runoffforecastingresearchcouplingquadraticfactorscreeninganddeeplearning |