基于RGCN-SA算法的海上浮标观测数据插补-海洋预报

基于RGCN-SA算法的海上浮标观测数据插补

单位：1. 国家海洋信息中心, 天津 300171;
2. 中国石油大学(华东)海洋与空间信息学院, 山东青岛 266580

分类号：P731.31

出版年·卷·期（页码）：2024·41·第五期（77-88）

摘要：

针对海洋观测数据的缺失问题，提出一种基于图卷积（GCN）和自注意力机制（SA）的残差网络插补模型（RGCN-SA），该模型由自注意力机制与图卷积构建，利用自注意力机制提取观测数据的时间依赖特征，通过图卷积获取不同位置浮标的空间依赖特征，并添加残差结构提高模型学习能力，结合自监督训练方式对模型进行训练，得到最终的海洋浮标数据插补模型。通过对比实验，证明该模型通过训练后能够有效获取浮标观测数据的时间与空间的关联特征，取得了比其他方法更好的插补效果。通过消融实验，证明模型的各个模块的有效性。

In this paper, a residual network imputation model based on graph convolution network (GCN) and self-attention mechanism (RGCN-SA) is proposed to solve the observational data missing problem. The model is constructed on self-attention mechanism and graph convolution. The self-attention mechanism is used to extract the time -dependent features of observational data, and the space-dependent features of buoys at different positions are obtained through graph convolution. Combined with the self-supervised training method, the model is trained and the final ocean data imputation model is obtained. Through comparative experiments, it is proved that the model can effectively obtain the temporal and spatial correlation features of buoy observations after training, and obtaina better imputation effect than other methods. The effectiveness of each module of the model is proved by the ablation experiment.

参考文献：

[1] 刘玉龙, 王国松, 侯敏, 等. 基于深度学习的海温观测数据质量控制应用研究[J]. 海洋通报, 2021, 40(3):283-291. LIU Y L, WANG G S, HOU M, et al. Quality control of sea temperature observation data using deep learning neural networks [J]. Marine Science Bulletin, 2021, 40(3):283-291.
[2] 郑婷婷, 于小宁, 陈旖旎. 基于海洋科技发展的观测监测档案资源动态整合与利用探究[J]. 北京档案, 2023(3):17-20. ZHENG T T, YU X N, CHEN Y N. Dynamic integration and utilization of observation and monitoring archive resources based on the development of marine science and technology[J]. Beijing Archives, 2023(3):17-20.
[3] ZHOU Y Z, SHI J S, STEIN R, et al. Missing data matter:an empirical evaluation of the impacts of missing EHR data in comparative effectiveness research[J]. Journal of the American Medical Informatics Association, 2023, 30(7):1246-1256.
[4] KANTARDZIC M. Data mining:concepts, models, methods, and algorithms[M]. Hoboken:John Wiley, 2011.
[5] AMIRI M, JENSEN R. Missing data imputation using fuzzy-rough methods[J]. Neurocomputing, 2016, 205:152-164.
[6] YANG K, LI J Z, WANG C K. Missing values estimation in microarray data with partial least squares regression[C]//6th International Conference on Computational Science. Reading:Springer, 2006:662-669.
[7] LI Y Y, PARKER L E. Nearest neighbor imputation using spatial– temporal correlations in wireless sensor networks[J]. Information Fusion, 2014, 15:64-79.
[8] AL-HELALI B, CHEN Q, XUE B, et al. A new imputation method based on genetic programming and weighted KNN for symbolic regression with incomplete data[J]. Soft Computing, 2021, 25(8):5993-6012.
[9] HASTIE T, MAZUMDER R, LEE J D, et al. Matrix completion and low-rank SVD via fast alternating least squares[J]. The Journal of Machine Learning Research, 2015, 16(1):3367-3402.
[10] JING X, LUO J G, WANG J M, et al. A multi-imputation method to deal with hydro-meteorological missing values by integrating chain equations and random forest[J]. Water Resources Management, 2022, 36(4):1159-1173.
[11] PAN H, YE Z W, HE Q Y, et al. Discrete missing data imputation using multilayer perceptron and momentum gradient descent[J]. Sensors, 2022, 22(15):5645.
[12] CAO W, WANG D, LI J, et al. BRITS:bidirectional recurrent imputation for time series[C]//Proceedings of the 32nd International Conference on Neural Information Processing Systems. Montréal:Curran Associates Inc., 2018:6776-6786.
[13] YOON J, JORDON J, VAN DER SCHAAR M. GAIN:missing data imputation using generative adversarial nets[C]//Proceedings of the 35th International Conference on Machine Learning. Stockholm:PMLR, 2018:5675-5684.
[14] DU W J, CÔTÉ D, LIU Y. SAITS:self-attention-based imputation for time series[J]. Expert Systems with Applications, 2023, 219:119619.
[15] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach:Curran Associates Inc., 2017:6000-6010.
[16] 贺琪, 曹万万, 黄冬梅, 等. 面向海洋时序数据异常模式发现的多视图协同可视分析[J]. 海洋通报, 2022, 41(6):619-629. HE Q, CAO W W, HUANG D M, et al. Multi-view collaborative visual analysis for anomaly detection of marine environment time series data[J]. Marine Science Bulletin, 2022, 41(6):619-629.

服务与反馈：

【文章下载】【发表评论】【查看评论】【加入收藏】