Temporal-relational graph neural network for nearshore seawater quality parameters multivariate multi-step prediction and correlation modelling

被引：0
作者：
Zhu, Qiguang [1 ]
Shen, Zhen [1 ]
Wu, Zhen [2 ]
Zhang, Hongbo [2 ]
Yuan, Junjun [1 ]
Chen, Ying [2 ]
机构：
[1] Yanshan Univ, Sch Informat Sci & Engn, Qinhuangdao, Hebei, Peoples R China
[2] Yanshan Univ, Sch Elect Engn, Qinhuangdao, Hebei, Peoples R China
来源：
EXPERT SYSTEMS WITH APPLICATIONS | 2025年 / 265卷
基金：
中国国家自然科学基金;
关键词：
Seawater quality parameters; Multivariate multi-step prediction; Graph neural network; Temporal convolution; Weighted adjacency matrix; Warning of algal bloom;
D O I：
10.1016/j.eswa.2024.126020
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
Accurate prediction of nearshore seawater quality parameters is important for nearshore environmental management. In this paper, a novel graph neural network model is proposed for realizing high-precision multivariate multi-step prediction of nearshore seawater quality parameters. By extracting and utilizing the correlations between parameters through a weighted adjacency matrix, this model addresses the issue of poor predictive performance in traditional deep neural networks, which stems from the difficulty in capturing and leveraging the relationships between parameters. The model's performance was evaluated using nearshore seawater quality parameter data from Beihai, Guangxi. It demonstrated satisfactory predictive performance, achieving an MSE of 0.0064 when forecasting data for the next 48 h, which outperformed several existing methods in comparative experiments. Validated the effectiveness of various structures in the model through ablation experiments. Further hyperparameter sensitivity analysis validated the reliability and generalizability of the model. The method also allows a brief modeling of the correlations between nearshore seawater quality parameters. Finally, the proposed model is deployed into the marine algal bloom analysis and early warning platform, which provides an important decision-making basis for algal bloom early warning and cause analysis.
引用
收藏
页数：16
相关论文
共 52 条
  • [1] Bai T., Tahmasebi P., Graph neural network for groundwater level forecasting, Journal of Hydrology, 616, (2023)
  • [2] Barzegar R., Aalami M.T., Adamowski J., Short-term water quality variable prediction using a hybrid CNN–LSTM deep learning model, Stochastic Environmental Research and Risk Assessment, 34, 2, pp. 415-433, (2020)
  • [3] Chang W., Li X., Dong H., Wang C., Zhao Z., Wang Y., Real-time prediction of ocean observation data based on transformer model, Paper presented at the Proceedings of the 2021 ACM International Conference on Intelligent Computing and its Emerging Applications, (2021)
  • [4] Chen Q., Cai C., Chen Y., Zhou X., Zhang D., Peng Y., TemproNet: A transformer-based deep learning model for seawater temperature prediction, Ocean Engineering, 293, (2024)
  • [5] Cohen I., Huang Y., Chen J., Benesty J., Benesty J., Chen J., Cohen I., Pearson correlation coefficient, Noise Reduction in Speech Processing, pp. 1-4, (2009)
  • [6] Cui B., Liu M., Li S., Jin Z., Zeng Y., Lin X., Deep learning methods for atmospheric PM2. 5 prediction: A comparative study of transformer and CNN-LSTM-attention, Atmospheric Pollution Research, 14, 9, (2023)
  • [7] Deng T., Chau K.-W., Duan H.-F., Machine learning based marine water quality prediction for coastal hydro-environment management, Journal of Environmental Management, 284, (2021)
  • [8] Du S., Li T., Yang Y., Horng S.-J., Multivariate time series forecasting via attention-based encoder–decoder framework, Neurocomputing, 388, pp. 269-279, (2020)
  • [9] Durr H.H., Laruelle G.G., van Kempen C.M., Slomp C.P., Meybeck M., Middelkoop H., Worldwide typology of nearshore coastal systems: Defining the estuarine filter of river inputs to the oceans, Estuaries and Coasts, 34, pp. 441-458, (2011)
  • [10] Fan J., Zhang K., Huang Y., Zhu Y., Chen B., Parallel spatio-temporal attention-based TCN for multivariate time series prediction, Neural Computing and Applications, 35, 18, pp. 13109-13118, (2023)
123456