Can I read A new water level prediction model based on ESMD−VMD−WSD−ESN on EtoBox?

A new water level prediction model based on ESMD−VMD−WSD−ESN by Xianqi Zhang; Haiyang Chen; Yihao Wen; Jinwen Shi; Yimeng Xiao is a scholarly article available to read on EtoBox.

What is A new water level prediction model based on ESMD−VMD−WSD−ESN about?

Water level prediction is critical for optimizing available water resources and sustainable use. We propose a new deep learning hybrid model for water level prediction. The model uses Extreme-point symmetric mode decomposition (ESMD), Variational Modal Decomposition(VMD), Wavelet Signal Denoising(WSD) and Echo State Network(ESN) to predict water level at four hydrological stations in the lower Yellow River. The model constructs the ESMD-VMD-WSD-ESN hybrid model by using the respective strengths to extract the most important features in the predictor variables. The prediction effectiveness of the combined quadratic decomposition model is explored by comparing it with the ESN, WSD-ESN and ESMD-WSD-ESN models. The results show that the quadratic decomposition-based combined model ESMD-VMD-WSD-ESN has the best prediction, ESMD-WSD-ESN is the second best, WSD-ESN and ESN are relatively poor. The Mean Absolute Error of the ESMD-VMD-WSD-ESN model is only 0.16 m, the Mean Absolute Percentage Error is 0.18% and the Nash-Sutcliffe efficiency coefficient reaches 0.91. In summary, the combined model of quadratic decomposition model has good applicability and stability in monthly water level pr

Author

Xianqi Zhang; Haiyang Chen; Yihao Wen; Jinwen Shi; Yimeng Xiao

Publisher

Springer Science and Business Media LLC

Published

2023

Language

EN