Research on Optimal Selection of Runoff Prediction Models Based on Coupled Machine Learning Methods DOI Creative Commons
Wei Xing,

M.M. Chen,

Yulin Zhou

и другие.

Research Square (Research Square), Год журнала: 2024, Номер unknown

Опубликована: Авг. 13, 2024

Abstract Runoff fluctuations under the influence of climate change and human activities present a significant challenge valuable application in constructing high-accuracy runoff prediction models. This study aims to address this by taking Wanzhou station Three Gorges Reservoir area as case optimize various The first selects artificial neural network (ANN) support vector machine (SVM) base Then, it evaluates from three time-series decomposition methods: Time-Varying Filter-based Empirical Mode Decomposition (TVF-EMD), Complete Ensemble with Adaptive Noise (CEEMDAN), Variational (VMD). Subsequently, these methods are coupled optimization algorithms, including Whale Optimization Algorithm (WOA), Grasshopper (GOA), Sparrow Search (SSA), construct hybrid results indicate that: (1) Among single models, Long Short-Term Memory (LSTM) model outperforms Backpropagation Neural Network (BP) SVM terms accuracy; (2) models show superior accuracy compared individual VMD-LSTM outperforming CEEMDAN-LSTM TVF-EMD-LSTM models; (3) learning VMD-SSA-LSTM achieves highest accuracy. Employing "decomposition-reconstruction" strategy combined robust algorithms enhances performance thereby significantly improving capabilities watershed hydrological

Язык: Английский

Research on tidal energy prediction method based on improved time-varying filter-empirical mode decomposition and confluent double-stream neural network DOI
Yi Huang, Guohui Li

Ocean Engineering, Год журнала: 2024, Номер 312, С. 119300 - 119300

Опубликована: Сен. 24, 2024

Язык: Английский

Процитировано

1
Research on optimal selection of runoff prediction models based on coupled machine learning methods DOI Creative Commons
Wei Xing,

Mengen Chen,

Yulin Zhou

и другие.

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Дек. 30, 2024

Язык: Английский

Процитировано

1
Stereoscopic Monitoring Methods for Flood Disasters Based on ICESat-2 and Sentinel-2 Data DOI Creative Commons

Yongqiang Cao,

Mengran Wang, Jiaqi Yao

и другие.

Remote Sensing, Год журнала: 2023, Номер 15(12), С. 3015 - 3015

Опубликована: Июнь 9, 2023

Climate change has led to an increased frequency of extreme precipitation events, resulting in damage from rainstorms and floods. Rapid efficient flood forecasting is crucial. However, traditional hydrological simulation methods that rely on site distribution are limited by the availability data cannot provide fast accurate monitoring information. Therefore, this study took event Huoqiu County 2020 as example proposes a three-dimensional method based active passive satellites, which provides effective information support for disaster prevention mitigation. The experimental results indicated following: (1) flood-inundated area was 704.1 km2, with Jiangtang Lake section Huaihe River southern part Chengdong being largest affected areas; (2) water levels ranged 15.36 m 17.11 m, 4–6 higher than original level. highest level areas were flat south Lake, north Chengxi having greatest increase; (3) depth primarily between 4 7 total storage capacity 2833.47 million m3, capacity; (4) rainstorm caused direct economic loss approximately CNY 7.5 billion population 91 thousand people. Three-dimensional floods comprehensively reflects inundation status can valuable prediction management.

Язык: Английский

Процитировано

3
Boosting flood routing prediction performance through a hybrid approach using empirical mode decomposition and neural networks: a case study of the Mera River in Ankara DOI Creative Commons
Okan Mert Katipoğlu, Metin Sarıgöl

Water Science & Technology Water Supply, Год журнала: 2023, Номер 23(11), С. 4403 - 4415

Опубликована: Окт. 31, 2023

Abstract Flood routing is vital in helping to reduce the impact of floods on people and communities by allowing timely appropriate responses. In this study, empirical mode decomposition (EMD) signal technique combined with cascade forward backpropagation neural network (CFBNN) feed-forward (FFBNN) machine learning (ML) techniques model 2014 Ankara, Mera River. The data are split order avoid underfitting overfitting problems algorithm. While establishing algorithm, 70% were divided into training, 15% testing validation. Graphical indicators statistical parameters used for analysis performance. As a result, EMD has been found improve performance ML models. addition, EMD-FFBNN hybrid showed most accurate estimation results flood calculation. study's outputs can assist designing control structures such as levees dams help risk.

Язык: Английский

Процитировано

3
Research on Optimal Selection of Runoff Prediction Models Based on Coupled Machine Learning Methods DOI Creative Commons
Wei Xing,

M.M. Chen,

Yulin Zhou

и другие.

Research Square (Research Square), Год журнала: 2024, Номер unknown

Опубликована: Авг. 13, 2024

Abstract Runoff fluctuations under the influence of climate change and human activities present a significant challenge valuable application in constructing high-accuracy runoff prediction models. This study aims to address this by taking Wanzhou station Three Gorges Reservoir area as case optimize various The first selects artificial neural network (ANN) support vector machine (SVM) base Then, it evaluates from three time-series decomposition methods: Time-Varying Filter-based Empirical Mode Decomposition (TVF-EMD), Complete Ensemble with Adaptive Noise (CEEMDAN), Variational (VMD). Subsequently, these methods are coupled optimization algorithms, including Whale Optimization Algorithm (WOA), Grasshopper (GOA), Sparrow Search (SSA), construct hybrid results indicate that: (1) Among single models, Long Short-Term Memory (LSTM) model outperforms Backpropagation Neural Network (BP) SVM terms accuracy; (2) models show superior accuracy compared individual VMD-LSTM outperforming CEEMDAN-LSTM TVF-EMD-LSTM models; (3) learning VMD-SSA-LSTM achieves highest accuracy. Employing "decomposition-reconstruction" strategy combined robust algorithms enhances performance thereby significantly improving capabilities watershed hydrological

Язык: Английский

Процитировано

0