Intelligent wave height prediction based on multitrunk fusion neural networks with bidirectional long short-term memory network
Year: 2025
Type: article
Abstract: This article introduces a multi-backbone neural network fusion model that integrates a Bidirectional Long Short-Term Memory network with residual connections, while incorporating convolutional layers, fully connected layers, and attention mechanisms. The model uses real-world marine environment data as its primary input, analyzes the nonlinear relationships between different backbones through a neural network framework, and jointly predicts future changes in ocean wave height. By training the model, the predicted values demonstrate a minimal error range compared to the actual values, highlighting its strong predictive capabilities and accurate results. Furthermore, the fusion of multi-backbone neural network models offers a novel optimization approach for predicting future marine environments. This proposed model provides innovative solutions for forecasting simulated marine environment data and intelligently predicting ship motion responses.
Source: Physics of Fluids
DOI: 10.1063/5.0251712
Authors: Chenhong Fan, Xingyi La, Franck Aurel Likeufack Mdemaya, Gang Xu
Institution: Jiangsu University of Science and Technology
Language: English
Subfield: Oceanography
Field: Earth and Planetary Sciences
Domain: Physical Sciences
Open Access status: hybrid
Award: G2433575884
OpenAlex ID: W4409212215